EMFT Exercises

Tobia Carozzi Anders Eriksson Bengt Lundborg Bo Thid Mattias Waldenvik
E LECTROMAGNETIC F IELD T HEORY E XERCISES
Draft version released 9th December 1999 at 19:47
Downloaded from http://www.plasma.uu.se/CED/Exercises
Companion volume to
E LECTROMAGNETIC F IELD T HEORY

by Bo Thid
E LECTROMAGNETIC F IELD T HEORY Exercises

Tobia Carozzi Anders Eriksson Bengt Lundborg Bo Thid Mattias Waldenvik
Department of Space and Plasma Physics Uppsala University and Swedish Institute of Space Physics Uppsala Division Sweden
Ipsum
A This book was typeset in LTEX 2 on an HP9000/700 series workstation and printed on an HP LaserJet 5000GN printer.
Copyright c 1998 by Bo Thid Uppsala, Sweden All rights reserved. Electromagnetic Field Theory Exercises ISBN X-XXX-XXXXX-X
C ONTENTS
Preface 1 Maxwells Equations

1.1 1.2 1.3 Coverage . . . . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . . . . Example 1.1 Macroscopic Maxwell equations . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix 1
1 1 1 1 2 4 4 5 5
Example 1.2 Maxwells equations in component form
Solution . . . . . . . . . . . . . . . . . . . . . . Example 1.3 The charge continuity equation . . . . Solution . . . . . . . . . . . . . . . . . . . . . .
2 Electromagnetic Potentials and Waves

2.1 2.2 2.3 Coverage . . . . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . . . . Example 2.1 The Aharonov-Bohm effect . . . Solution . . . . . . . . . . . . . . . . . . . Example 2.2 Invent your own gauge . . . . . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
9 9 9 9 10 11 11 13 13 15 15
Example 2.3 Fourier transform of Maxwells equations
Solution . . . . . . . . . . . . . . . . . . . . . . . Example 2.4 Simple dispersion relation . . . . . . . . Solution . . . . . . . . . . . . . . . . . . . . . . .
3 Relativistic Electrodynamics
3.1 Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
17 i
ii
3.2 3.3
Formulae used . . . . . . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . . . . . . Example 3.1 Covariance of Maxwells equations . Solution . . . . . . . . . . . . . . . . . . . . .
17 18 18 18 Example 3.2 Invariant quantities constructed from the eld tensor 20 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Example 3.3 Covariant formulation of common electrodynamics formulas . . . . . . . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 3.4 Fields from uniformly moving charge via Lorentz transformation . . . . . . . . . . . . . . . . . . .
21 21 23 23
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Lagrangian and Hamiltonian Electrodynamics

4.1 4.2 4.3 Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . . . . . . . . . . . . Example 4.1 Canonical quantities for a particle in an EM eld . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . Example 4.2 Gauge invariance of the Lagrangian density . . . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
27 27 28 28 28 29 29
5 Electromagnetic Energy, Momentum and Stress

5.1 5.2 5.3 Coverage . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . Example 5.1 EM quantities potpourri . . Solution . . . . . . . . . . . . . . . . Example 5.2 Classical electron radius . Solution . . . . . . . . . . . . . . . . Example 5.3 Solar sailing . . . . . . . Solution . . . . . . . . . . . . . . . . . . . . . . . . . Example 5.4 Magnetic pressure on the earth . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
31 31 32 32 32 35 35 37 37 39 39
6 Radiation from Extended Sources

6.1 6.2 6.3 Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . . . . . . . . . . . .
41
41 41 42 Example 6.1 Instantaneous current in an innitely long conductor 42

iii
Solution . . . . . . . . . . . . . . . . . Example 6.2 Multiple half-wave antenna . Solution . . . . . . . . . . . . . . . . . Example 6.3 Travelling wave antenna . . . Solution . . . . . . . . . . . . . . . . . Example 6.4 Microwave link design . . . Solution . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
42 47 47 50 50 51 51
7 Multipole Radiation
7.1 7.2 7.3 Coverage . . . . . . . . . . . . . . . . . . . Formulae used . . . . . . . . . . . . . . . . . Solved examples . . . . . . . . . . . . . . . Example 7.1 Rotating Electric Dipole . Solution . . . . . . . . . . . . . . . . Example 7.2 Rotating multipole . . . . Solution . . . . . . . . . . . . . . . . Example 7.3 Atomic radiation . . . . . Solution . . . . . . . . . . . . . . . . Example 7.4 Classical Positronium . . . Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
53 53 54 54 54 56 56 58 58 59 59
8 Radiation from Moving Point Charges

Example 8.1 Poynting vector from a charge in uniform motion
63
63 63 64 64 64 66 66 67 67 69 69
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 8.2 Synchrotron radiation perpendicular to the acceleration . . . . . . . . . . . . . . . . . . . . . .
Solution . . . . . . . . . . . . . . . . . Example 8.3 The Larmor formula . . . . Solution . . . . . . . . . . . . . . . . . Example 8.4 Vavilov-Cerenkov emission . Solution . . . . . . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
9 Radiation from Accelerated Particles

71
71 71 72
iv
Example 9.1 Motion of charged particles in homogeneous static EM elds . . . . . . . . . . . . . . . . . . . . .
Solution . .
Example 9.2
Solution . .
Example 9.3
Solution . .
Example 9.4
Solution . .
72 . . . . . . . . . . . . . . . . . . . . . . . . . 72 Radiative reaction force from conservation of energy 74 . . . . . . . . . . . . . . . . . . . . . . . . . 74 Radiation and particle energy in a synchrotron . . 77 . . . . . . . . . . . . . . . . . . . . . . . . . 77 Radiation loss of an accelerated charged particle . 79 . . . . . . . . . . . . . . . . . . . . . . . . . 79
F Formulae
F.1 The Electromagnetic Field . . . . . . . . . . . . . . . . . . . F.1.1 Maxwells equations . . . . . . . . . . . . . . . . . . Constitutive relations . . . . . . . . . . . . . . . . . . F.1.2 Fields and potentials . . . . . . . . . . . . . . . . . . Vector and scalar potentials . . . . . . . . . . . . . . Lorentz gauge condition in vacuum . . . . . . . . . . F.1.3 Force and energy . . . . . . . . . . . . . . . . . . . . Poyntings vector . . . . . . . . . . . . . . . . . . . . Maxwells stress tensor . . . . . . . . . . . . . . . . . Electromagnetic Radiation . . . . . . . . . . . . . . . . . . . F.2.1 Relationship between the eld vectors in a plane wave F.2.2 The far elds from an extended source distribution . . F.2.3 The far elds from an electric dipole . . . . . . . . . . F.2.4 The far elds from a magnetic dipole . . . . . . . . . F.2.5 The far elds from an electric quadrupole . . . . . . . F.2.6 The elds from a point charge in arbitrary motion . . . F.2.7 The elds from a point charge in uniform motion . . . Special Relativity . . . . . . . . . . . . . . . . . . . . . . . . F.3.1 Metric tensor . . . . . . . . . . . . . . . . . . . . . . F.3.2 Covariant and contravariant four-vectors . . . . . . . . F.3.3 Lorentz transformation of a four-vector . . . . . . . . F.3.4 Invariant line element . . . . . . . . . . . . . . . . . . F.3.5 Four-velocity . . . . . . . . . . . . . . . . . . . . . . F.3.6 Four-momentum . . . . . . . . . . . . . . . . . . . . F.3.7 Four-current density . . . . . . . . . . . . . . . . . . F.3.8 Four-potential . . . . . . . . . . . . . . . . . . . . . . F.3.9 Field tensor . . . . . . . . . . . . . . . . . . . . . . . Vector Relations . . . . . . . . . . . . . . . . . . . . . . . . . F.4.1 Spherical polar coordinates . . . . . . . . . . . . . . .
83
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 83 83 84 84 84 84 84 84 84 84 84 85 85 85 85 86 86 86 86 86 87 87 87 87 87 87 87 88
F.2
F.3
F.4
F.4.2
Base vectors . . . . . Directed line element . Solid angle element . . Directed area element Volume element . . . Vector formulae . . . . General relations . . . Special relations . . . Integral relations . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
88 88 88 88 88 89 89 90 91
vi
L IST
OF
F IGURES
9.1
Motion of a charge in an electric and a magnetic eld . . . . . . .
6.1 6.2 6.3
The turn-on of a linear current at t 0 . . . . . . . . . . . . . . Snapshots of the eld . . . . . . . . . . . . . . . . . . . . . . . . Multiple half-wave antenna standing current . . . . . . . . . . . .
43 44 47 74
vii
viii
P REFACE
This is a companion volume to the book Electromagnetic Field Theory by Bo Thid. The problems and their solutions were created by the co-authors who all have taught this course or its predecessor. It should be noted that this is a preliminary draft version but it is being corrected and expanded with time. Uppsala, Sweden December, 1999 B. T.
ix
P REFACE
L ESSON 1
Maxwells Equations
1.1 Coverage
In this lesson we examine Maxwells equations, the cornerstone of electrodynamics. We start by practising our math skill, refreshing our knowledge of vector analysis in vector form and in component form.
1.2
Formulae used
!
B E B
1.3

Solved examples
E XAMPLE 1.1 (1.2a) (1.2b) (1.2c) 1 E c2 t (1.2d)
M ACROSCOPIC M AXWELL EQUATIONS The most fundamental form of Maxwells equations is
"
0 j
E B
0 0 B t
0 j

0 0 B t
(1.1a) (1.1b) (1.1c) 1 E c2 t (1.1d)
L ESSON 1. M AXWELL S E QUATIONS
sometimes known as the microscopic Maxwell equations or the Maxwell-Lorentz equations. In the presence of a medium, these equations are still true, but it may sometimes be convenient to separate the sources of the elds (the charge and current densities) into an induced part, due to the response of the medium to the electromagnetic elds, and an extraneous, due to free charges and currents not caused by the material properties. One then writes
The electric and magnetic properties of the material are often described by the electric polarisation P (SI unit: C/m2 ) and the magnetisation M (SI unit: A/m). In terms of these, the induced sources are described by
To fully describe a certain situation, one also needs constitutive relations telling how P and M depends on E and B. These are generally empirical relations, different for different media. Show that by introducing the elds
the two Maxwell equations containing source terms (1.2a) and (??) reduce to
known as the macroscopic Maxwell equations.
Solution
If we insert
and
"
"
jind ind
jext ext
"
ext
jext
"

D H
0 E P B 0 M
"
$
!
jind ind
P t P
#
"
"

jind ind
jext ext
(1.3) (1.4)
(1.5) (1.6)
(1.7) (1.8)
(1.9)
D t
(1.10) (1.11)
(1.12) (1.13)
1.3. S OLVED
EXAMPLES
into
we get
which can be rewritten as
ext
Now by introducing the D and the H elds such that
the Maxwell equations become
The reason these equations are known as macroscopic are that the material properties described by P and M generally are average quantities, not considering the atomic properties of matter. Thus E and D get the character of averages, not including details around single atoms etc. However, there is nothing in principle preventing us from using largescale averages of E and B, or even to use atomic-scale calculated D and H although this is a rather useless procedure, so the nomenclature microscopic/macroscopic is somewhat misleading. The inherent difference lies in how a material is treated, not in the spatial scales.
E ND
OF EXAMPLE
1.1
ext
"

"
0 E P B M 0
jext
D t
&
" %
"
1&
0)
B M 0 0 E P
jext
&
"

1 0 ext
P t
" &
"
"
0 jext
#
"
!
B
ind
0 j
"

$
' (
jind
#
P t
(1.14) (1.15)
1 E c2 t

(1.16) (1.17) (1.18)
P t
1 E c2 t
(1.19) (1.20)
0 E
(1.21) (1.22)
(1.23) (1.24)
(1.25) (1.26) QED
E XAMPLE 1.2
M AXWELL S EQUATIONS IN COMPONENT FORM Express Maxwells equations in component form.
Solution
Maxwells equations in vector form are written:
In these equations, E, B, and j are vectors, while is a scalar. Even though all the equations contain vectors, only the latter pair are true vector equations in the sense that the equations themselves have several components. When going to component notation, all scalar quantities are of course left as they are. Vector quantities, for example E, can always be expanded as E E j x j, 3 1 E j x j j where the last step assumes Einsteins summation convention: if an index appears twice in the same term, it is to be summed over. Such an index is called a summation index. Indices which only appear once are known as free indices, and are not to be summed over. What ak bk , symbol is used for a summation index is immaterial: it is always true that ai bi since both these expressions mean a1 b1 a2 b2 a3 b3 a b. On the other hand, the expression ai ak is in general not true or even meaningful, unless i k or if a is the null vector. The three E j are the components of the vector E in the coordinate system set by the three unit vectors x j . The E j are real numbers, while the x j are vectors, i.e. geometrical objects. Remember that though they are real numbers, the E j are not scalars. Vector equations are transformed into component form by forming the scalar product of both sides with the same unit vector. Let us go into ridiculous detail in a very simple case:
This is of course unnecessarily tedious algebra for an obvious result, but by using this careful procedure, we are certain to get the correct answer: the free index in the resulting equation necessarily comes out the same on both sides. Even if one does not follow this complicated way always, one should to some extent at least think in those terms. Nabla operations are translated into component form as follows:
&
%
G xk G j x j xk G j jk Gk
H xk Hi x i x k Hi ik Hk
"
"
"
!
0 j
&
$ $
E B
0 0 B t
(1.27) (1.28) (1.29) 1 E c2 t (1.30)
(1.31) (1.32) (1.33) (1.34) (1.35)
1.3. S OLVED
EXAMPLES
5
xi Vi xi V i jk k xj
where V is a vector eld and is a scalar eld. Remember that in vector valued equations such as Ampres and Faradays laws, one must be careful to make sure that the free index on the left hand side of the equation is the same as the free index on the right hand side of the equation. As said above, an equation of the form Ai B j is almost invariably in error! With these things in mind we can now write Maxwells equations as
T HE CHARGE CONTINUITY EQUATION Derive the continuity equation for charge density from Maxwells equations using (a) vector notation and (b) component notation. Compare the usefulness of the two systems of notations. Also, discuss the physical meaning of the charge continuity equation.
Solution
Vector notation In vector notation, a derivation of the continuity equation for charge
looks like this: Compute
t E t
in two ways:
1. Apply
to Gausss law:
6&
$
1 0 t
(1.43)
"
5 !
"
0 j
1 E c2 t
i jk
Bk xj
!
5 !
!
B t
5 !
5 !
5 !

$
#
5 ! 5 !
xi V V xi i i jk x i V xj k
xi
(1.36) (1.37) (1.38)
$
Ei xi 0 Bi 0 xi E i jk k xj
(1.39) (1.40)
B t i
1 Ei c2 t
(1.41) (1.42)
0 ji
E ND
OF EXAMPLE
1.2
E XAMPLE 1.3
2. Take the divergence of the Ampre-Maxwell law: B
Comparison shows that
Component notation In component notation, a derivation of the continuity equation

for charge looks like this: Compute
xi t Ei t
in two ways:
1. Take
of Gausss law:
2. Take the divergence of the Ampre-Maxwell law:
Comparison shows that
"
ji xi
!
Ei t xi
1 j 0 x i i
Use that the relation i jk Ai A j
0 is valid also if Ai
"
xi
i jk
Ei t xi
1 0 t
Bk xj
$
"

!
E t
1 0
j xi i

j
8 6&
Use
0 and 0 0 c2
"
1: (1.45)

1 c2
E t

6&
$
% % 7 9 @ 5 !
(1.44)
(1.46)
(1.47)
1 Ei c2 xi t
xi ,
(1.48)
and that 0 0 c2
1:
(1.49)
(1.50)
1.3. S OLVED
EXAMPLES
Comparing the two notation systems We notice a few points in the derivations
above:
Interpreting the continuity equation The equation

j 0 (1.51) t is known as a continuity equation. Why? Well, integrate the continuity equation over some volume V bounded by the surface S. By using Gausss theorem, we nd that
dQ j dS d3x j d3x (1.52) dt t V V S which says that the change in the total charge in the volume is due to the net inow of electric current through the boundary surface S. Hence, the continuity equation is the eld theory formulation of the physical law of charge conservation.
F !
7
F !
E
F 7
"
D D D D
It is sometimes difcult to see what one is calculating in the component system. The vector system with div, curl etc. may be closer to the physics, or at least to our picture of it. In the vector notation system, we sometimes need to keep some vector formulas in memory or to consult a math handbook, while with the component system you need only the denitions of i jk and i j . Although not seen here, the component system of notation is more explicit (read unambiguous) when dealing with tensors of higher rank, for which vector notation becomes cumbersome. The vector notation system is independent of coordinate system, i.e., is in any coordinate system, while in the component notation, the components depend on the unit vectors chosen.
E ND
OF EXAMPLE
1.3
L ESSON 2
Electromagnetic Potentials and Waves

2.1 Coverage
Here we study the vector and scalar potentials A and and the concept of gauge transformation. One of the most important physical manifestation of Maxwells equations is the EM wave. Seen as wave equations, the Maxwell equations can be reduced to algebraic equations via the Fourier transform and the physics is contained in so-called dispersion relations which set the kinematic restrictions on the elds.
2.2
E
Formulae used

A
A t
2.3

T HE A HARONOV-B OHM
Consider the magnetic eld given in cylindrical coordinates,
Q& P P Q& P P
I %
R %
Br Br
H G
Solved examples
EFFECT
E XAMPLE 2.1
r0 z r0 z
B z 0
(2.1) (2.2)
10
L ESSON 2. E LECTROMAGNETIC P OTENTIALS
AND
WAVES
Determine the vector potential A that generated this magnetic eld.
Solution
A interesting question in electrodynamics is whether the EM potentials and A are more than mathematical tools, and alternatives to the Maxwell equations, from which we can derive the EM elds. Could it be that the potentials and not Maxwells equations are more fundamental? Although the ultimate answer to these questions is somewhat metaphysical, it is exactly these questions that make the Aharonov-Bohm effect. Before we discuss this effect let us calculate the vector eld from the given magnetic eld. The equations connecting the potentials with the elds are
In this problem we see that we have no boundary conditions for the potentials. Also, let us use the gauge 0. This problem naturally divides into two parts: the part within the magnetic eld and the part outside the magnetic eld. Let us start with the interior part:
A r z . Examining the The rst equation tells us that A is time independent so A other three we nd that there is no dependence on or z either so A A r . All that remains is 1 rA B (2.6) r r Integrating this equation we nd that
Moving to the outer problem, we see that the only difference compared with the inner problem is that B 0 so that we must consider
&
1 rA r r
Br 2
& %
& P P %
)
&
&
'
1 r
A t A 1 Az r z Ar Az z r rA Ar r
A t
# !
(2.3) (2.4)
0 0 0 B
(2.5a) (2.5b) (2.5c) (2.5d)
(2.7)
(2.8)
2.3. S OLVED
EXAMPLES
11
This time integration leads to C (2.9) r If we demand continuity for the function A over all space we nd by comparing with (2.7) the arbitrary constant C and can write in outer solution as
Now in electrodynamics (read: in this course) the only measurable quantities are the elds. So the situation above, where we have a region in which the magnetic eld is zero but the potential is non-zero has no measurable consequence in classical electrodynamics. In quantum mechanics however, the Aharonov-Bohm effect shows that this situation does have a measurable consequence. Namely, when letting charged particles go around this magnetic eld (the particles are do not enter the magnetic eld region because of a impenetrable wall) the energy spectrum of the particles after passing the cylinder will have changed, even though there is no magnetic eld along their path. The interpretation is that the potential is a more fundamental quantity than the eld.
I NVENT
YOUR OWN GAUGE
Name some common gauge conditions and discuss the usefulness of each one. Then invent your own gauge and verify that it is indeed a gauge condition.
Solution
Background The Maxwell equations that do not contain source terms can be solved by using the vector potential A and the scalar potential , dened through the relations
Assuming linear, isotropic and homogeneous media, we can use the constitutive relations E H B , and j E j (where j is the free current density arising from D other sources than conductivity) and denitions of the scalar and vector potentials in the remaining two Maxwell equations and nd that
! W)
"
"
2 A
V'
2A A 2 t t
!
"
A t
"
U
!
(2.11)
A t
(2.12)
(2.13)
(2.14)
#
2 Br0 2r
0!
(2.10)
E ND
OF EXAMPLE
2.1
E XAMPLE 2.2
12
AND
WAVES
These equations are used to determine A and from the source terms. And once we have found A and it is straight forward to derive the E and B elds from (2.11) and (2.12). The denitions of the scalar and vector potentials are not enough to make A and unique, i.e. , if one is given A and then (2.11) and (2.12) determine B and E, but if one is given B and E there many ways of choosing A and . This can be seen through the fact that A and can be transformed according to
where is an arbitrary scalar eld, but the B and E elds do not change. This kind of transformation is called a gauge transformation and the fact that gauge transformations do not affect the physically observable elds is known as gauge invariance.
Gauge conditions The ambiguity in the denitions of A and can be used to introduce a gauge condition. In other words, since the denitions (2.11) and (2.12) do not completely dene A and we are free to add certain conditions. Some common gauge conditions are
A
The Lorentz gauge is the most commonly used gauge for time-varying elds. In this case (2.13) and (2.14) reduce to
So the Lorentz transform decouples (2.13) and (2.14) and puts and A on equal footing. Furthermore, the resulting equations are manifestly covariant.
XYX
"
1 2A c2 t 2
In the temporal gauge one discards the scalar potential by setting (2.13) and (2.14) reduce to
!
!
!
2 A
!
t t
2 t2 2 2 A t
0. In this gauge
The Coulomb gauge is most useful when dealing with static elds. Using (2.13) and (2.14, for static elds, reduces to
$
"
"

Coulomb gauge Lorentz gauge Temporal gauge

T
"
T
(2.15) (2.16)
0 0 0 A 0 then
(2.17) (2.18)
' '
(2.19) (2.20)
(2.21)
2.3. S OLVED
EXAMPLES
13
Thus the single vector A describes both E and B in the temporal gauge.
How to invent your own gauge Gauges other than Coulomb, Lorentz and the temporal mentioned above are rarely used in introductory literature in Electrodynamics, but it is instructive to consider what constitutes a gauge condition by constructing ones own gauge. Of course, a gauge condition is at least a scalar equation containing at least one of the components of A or . Once you have an equation that you think might be a gauge, it must be veried that it is a gauge. To verify that a condition is a gauge condition it is sufcient to show that any given set of A and can be made to satisfy your condition. This is done through gauge transformations. So given a A and a which satisfy the physical conditions through (2.13) and (2.14) we try to see if it is possible (at least in principle) to nd a gauge transformation to some new potential A and , which satisfy your condition.
F OURIER TRANSFORM OF M AXWELL S EQUATIONS Fourier transform Maxwells Equation. Use the Fourier version of Maxwells equations to investigate the possibility of waves that do not propagate energy; such waves are called static waves.
Solution
Maxwells equations contain only linear operators in time and space. This makes it easy to Fourier transform them. By transforming them we get simple algebraic equations instead of differential equations. Furthermore, the Fourier transformed Maxwell equations are useful when working with waves or time-varying elds, especially since the response function, i.e. the dielectric function, is in many case more fundamentally described as a function of angular frequency than length x. To perform this derivation we need formulas on how to translate the operators and t in Maxwells equations. The Fourier transform in time, is dened by

and the Fourier transform in space, is dened analogously
and so a combined spatial and time Fourier transform becomes
If we apply the last formula on
E we get
& P % e

b 0a& P % F
F k
dt d3x e
b c d b
& %
b 76& % F 8
Ak
d3x e
ik x
& %
c b
b 7a& % F 8
dt ei t f t
(2.22)
Ax
(2.23)
i k x t
F t x
(2.24)
3
E XAMPLE 2.3
7 `
E ND
OF EXAMPLE
2.2
14
AND
WAVES
iki
dt d3x Ei t x
iki Ei k
ik E k
where we have used partial integration. For
ii jk k j Ek k ei
where we have once again used partial integration. One may proceed analogously for tE t x . Trivially, one gets similar equations for the transformation of the D, H and B elds. Thus we have found that
where V t x is an arbitrary eld and denotes here Fourier transform. These transformation rules are easy to remember by saying that roughly the Fourier transform of is ik and the Fourier transform of t is i . Now we can use (2.27a), (2.27b) and (2.27c) on Maxwells equations. We then get, after some simple trimming
& P %
Q& & P % ! Q& ! & P % Q& & P % Q&
P % P % P %
P %
k ik k k
E H D B
k k k k
B k k j i D k i k 0
5 P % & 51& P % 5 1& P %
Vt x Vt x V t x t
ik V k ik V k i V k
(2.27a) (2.27b) (2.27c)
(2.28a) (2.28b) (2.28c) (2.28d)
dt Ei
dt d3x
P % b c d b &
t F ! e y
i k x t
& P % 4 r sq p AA b c d b sxA w
dt d3x e
i k x t
i jk
Ej t x ei xk
ii jk k j Ek t x e
i k x t
& P % & P %
& P % ! & P % & P %
b c d b
F v5
dt d3x e
# e
i k x t
Et
t i
E we get
& u& P %
!%
F !
dt
xe
g he
& P % 6& P % & P % F 4 r sq p & P % f F b c d b
Ei t
i k x t
dt d3x e
& P %
dt d3x iki Ei t x (2.25) ei (2.26)
b c d b
b c d b
b F b G F F F 5
dt d3x e
i k x t
e
i k x t
b c d b
& P %
$
& P %
$
Ei t x xi
2.3. S OLVED
EXAMPLES
15
where we have dropped the notation. These are the Fourier versions of Maxwells equation. As an example of the use of the Fourier transformed Maxwells equations let us derive static waves. Static waves are one possible oscillation mode for the E and H elds. Lets say that we have a mode such that the E E eld is oscillating at 0 and that it has a k k 0 which is parallel to the electric eld, so k E . From (2.28a) this implies that
So, we see that S E H 0 trivially. The lesson here is that you can have time-varying elds that do not transmit energy! These waves are also called longitudinal waves for obvious reasons.
S IMPLE
DISPERSION RELATION
If a progressive wave is travelling in a linear, isotropic, homogenous, nonconducting dielectric medium with dielectricity parameter and permeability , what is the dispersion relation? And what is the group velocity in this case? Also, what is the dispersion relation in a conducting medium?
Solution
A dispersion relation is a relation between and k, usually something like From this one can solve for which is then a function of k1 k2 k3 . For isotropic media then will be a function of k k only. A dispersion relation determines what modes (i.e. what combinations of k and ) are possible. The dispersion relation is derivable in principal once one has explicit knowledge of the dielectricity function (or response function) for the medium in question. The two vector equations in Maxwells equations are D k
(2.31)
so for a progressive wave characterized by and k propagating in a linear, isotropic, homogeneous medium with 0 (so j E 0), then these equations give
"
!
B t D j t
(2.32) (2.33)
&
6& P %
E ND
OF EXAMPLE
(2.29) (2.30)
2.3
#
E XAMPLE 2.4
16
AND
WAVES
Operating on (2.34) with k
0 Progressive wave!
Since E is not assumed to be zero then
The group velocity in this case is vg uk u k k so in this simple case the group velocity is the same as the phase velocity.
For the case of a conducting medium, in which j E, the two vector equations applied on a wave which at rst resembles the progressive wave we used above gives
Combining these two equation as done previously, we get
So that
If
4u2 k2 2 2 0 then we are back again to the previous problem as can be veried.
0
2u2 k
i 1 4u2 k2 2 2 and the group velocity is
2 2
"
uk
2 2
!
k2 E
i E
!
E B
B
i E
2 E
E 0
where we have identied the phase velocity u
6&
0
0
k2
2 0 k2 k ku .
&
k2
2 E
!
6&
& t rsq % p %
k E k E
k B k k2 E
!
8
E B
B E
and then using (2.35) we get a single vector equation:
(2.34) (2.35)
2 E
(2.36) (2.37) (2.38)
9 9 9 9 9
(2.39) (2.40) (2.41)
(2.42)
(2.43) (2.44)
(2.45) (2.46)
(2.47)
(2.48)
E ND
OF EXAMPLE
2.4
L ESSON 3
Relativistic Electrodynamics
3.1 Coverage
We examine the covariant formulation of electrodynamics. We take up the concept of 4-tensors and give examples of these. Also, show how 4-tensors are manipulated. We discuss the group of Lorentz transformations in the context of electrodynamics.
3.2
Formulae used
(3.1)
A Lorentz boost in the 3-direction

L
The eld tensor (components are 0, 1, 2, 3)
0 E1 E2 E3
E1 0 cB3 cB2
E2 cB3 0 cB1
E3 cB2 cB1 0
0 0
0 1 0 0
0 0 1 0
0 0
(3.2) 17
18
L ESSON 3. R ELATIVISTIC E LECTRODYNAMICS
3.3
E XAMPLE 3.1
Solved examples
C OVARIANCE OF M AXWELL S EQUATIONS Discuss the covariance of Maxwells equations by showing that the wave equation for electromagnetic elds is invariant with respect to Lorentz transformations but not Galilean transformations.
Solution
The dAlembert operator 1 2 2 (3.3) c2 t 2 is a fundamental operator in electrodynamics. The dynamics of EM elds is completely described using the dAlembertian. Galilean transformations, even though closest to our intuitive picture of the fabric of space-time, does not leave the dAlembertian invariant. A Galilean transformation is simply
where the origin of the primed system is moving relative the unprimed along the 3-direction with velocity v. Now we introduce this transformation by expanding each differential in the unprimed coordinate system in terms of the differential in the primed system by using the chain rule of derivation, i.e. we evaluate x x x x , so
"
"
"
x3
t x3 t x3
x1 x3 x1
x2 x3 x2
x3 x3 x3
"
"
"
x2
t x2 t x2
x1 x1 x1
x2 x2 x2
x3 x2 x3
"
"
"
x1
t x1 t x1
x1 x1 x1
x2 x1 x2
x3 x1 x3
&
T
%
T
T
fT T
x1 x2 x3 t
x1 x2 x3
vt
(3.4)
(3.5) (3.6)
(3.7) (3.8)
(3.9) (3.10)
3.3. S OLVED
EXAMPLES
19
and so
but
and so
where we have used the fact that the operators t and x3 commute. Thus we have found that
Which obviously does not have the same form as the dAlembertian in the unprimed system! Let us do the same calculations for the case of a Lorentz transformation; more specically we consider a boost along the 3 axis which is given by
where
(remember that runs over 0, 1, 2, 3). Since and depend on v The 4-gradient, x transforms as
so
In other words we have found that the dAlembertian is invariant under Lorentz boosts.
E ND
OF EXAMPLE
g g
&
% !
&
2 2 2 0
2 1
g g
%
g L L
2 2
2 2 2 3
g
"
g
g
x x
L x x
hjjk i g
0 0
0 1 0 0
0 0 1 0
0 0
mm sl
g
L x
L x
"
1 2 c2 t 2
v2 2 c2 x32
2v 2 c2 t x3
"
6&
%& T
T
%
2 t2
x3
x3
2 2 2 v2 2 2v t 2 x3 t x3
"
"
"
t x1 t t t x1 v t x3
x2 t x2
2 x2 1
2 x12 ;
2 x2 2
2 x22 ;
2 x2 3
2 x32
(3.11)
x3 t x3
(3.12) (3.13)
8
(3.14)
(3.15)
(3.16)
(3.17)
(3.18)
(3.19)
3.1
20
E XAMPLE 3.2
I NVARIANT QUANTITIES CONSTRUCTED FROM THE FIELD TENSOR 1 F of the eld tensor F . What quantities Construct the dual tensor 2 constructed solely with the eld tensor and its dual tensor, are invariant under Lorentz transformations? Having found these quantities you should be able to answer the questions: can a purely electric eld in one inertial system be seen as a purely magnetic eld in another?
and, can a progressive wave be seen as a purely electric or a purely magnetic eld in an inertial system?
Solution
The dual tensor of F is given by
where
and g
We determine rst the eld tensor with two covariant indices through the formula
so
From the formula for the dual tensor we see that it is a 4-tensor since F is a four tensor and is easily shown to be an invariant under orthogonal transforms for which the Lorentz transform is a special case. What can we create from and F which is invariant under Lorentz transformations? We consider the obviously invariant quantities
We see that the dual tensor can be obtained from F by putting E
0 cB1 cB2 cB3
cB1 0 E3 E2
cB2 E1 0 E1
cB3 E2 E1 0
mm l
hjj i
g g F
! hjjk ! i
0 E1 E2 E3
E1 0 cB3 cB2
E2 cB3 0 cB1
E3 cB2 cB1 0
mm sl
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
mm l
hjj i
0 E1 E2 E3
E1 0 cB3 cB2
E2 cB3 0 cB1
E3 cB2 cB1 0
mm l
hjj i
1 F 2
1 g g F 2
(3.20)
(3.21)
(3.22)
(3.23)
(3.24)
cB and B
E c.
3.3. S OLVED
EXAMPLES
21
F and F F . F F
Relation of EM elds in different inertial systems Now that we know that E B and
E 2 c2 B2 are Lorentz invariant scalars, let see what they say about EM elds in different inertial systems. Let us say that X E B and Y E 2 c2 B2 . All inertial systems must have the same value for X and Y . A purely electric eld in one inertial system means that B 0, so X 0 and Y 0. A purely magnetic eld would mean that E 0, so X 0 but Y 0. In other words it does not seem that a purely electric eld can be a purely magnetic eld in any inertial system. For a progressive wave E B so X 0 and in a purely electric or a purely magnetic eld 0 X 0 also, but for a progressive wave E cB so Y 0 and if the other system has E or B 0 then Y 0 force both the elds to be zero. So this is not possible.
C OVARIANT FORMULATION OF COMMON ELECTRODYNAMICS FORMULAS Put the following well know formulas into a manifestly covariant form
The continuity equation
Lorentz force
The inhomogeneous Maxwell equations
The homogenous Maxwell equations
Solution
The Methodology To construct manifestly covariant formulas we have at our disposal the following building blocks:
Also we have the 4-gradient
&
P & P & P & P & P P & P
&
% % % % % % %
an event 4-velocity 4-momentum wave 4-vector 4-current density 4-potential 4-force
x u p k J A F
ct x , c v , E c p, c k, c j , c A, v F c F
E ND
OF EXAMPLE
fT
3.2
qT
This means that E B and E 2
&
!
4cE B E2
2 c2 B2
fT
(3.25) (3.26) c2 B2 are Lorentz invariant scalars.
E XAMPLE 3.3
22
as our operator building block and also the second rank 4-tensor
Observe that we use indices which run 0, 1, 2, 3 where the 0-component is time-like component, there is also the system where indices run 1, 2, 3, 4 and the 4-component is the time-like component. Beware! A sufcient condition to formulate covariant electrodynamic formulas is that we make our formulas by combine the above 4-vectors. To make sure we have a covariant form we take outer product (i.e. simply combine the tensors so that all the indices are free) and then perform zero or more contractions, i.e. equate two indices and sum over this index (notationally this means we create a repeated index). In the notation we use here contractions must be between a contravariant (upper) index and a covariant (lower) index. One can always raise or lower a index by including a metric tensor g . On top of this sufcient condition, we will need to use our knowledge of the formulas we will try to make covariant, to accomplish our goal.
The continuity equation We know that the continuity equation is a differential equation which includes the charge density and the current density and that it is a scalar equation. This leads us to calculate the contraction of the outer product between the 4-gradient and the 4-current J
This is covariant version of the continuity equation, thus in space-time the continuity equation is simply stated as the 4-current density is divergence-free!
Lorentz force We know that the left hand side of Lorentz force equation is a 3-force.
Obviously we should use the covariant 4-vector force instead. And on the right hand side of th Lorentz equation is a 3-vector quantity involving charge density and current density and the E and B elds. The EM elds are of course contained in the eld tensor F . To get a vector quantity from F and J we contract these so our guess is
Inhomogeneous Maxwell equations The inhomogeneous Maxwell may be written

as the 4-divergence of the eld tensor
F J
hjj i
eld tensor
0 E1 E2 E3
E1 0 cB3 cB2
E2 cB3 0 cB1
E3 cB2 cB1 0
mm l
&
(3.27) (3.28) (3.29)
P
%
4-gradient
ct x
3.3. S OLVED
EXAMPLES
23
Homogeneous Maxwell equations The homogenous Maxwell equations are written

most compactly using the dual tensor of the eld tensor. Using the dual tensor we have
F IELDS FROM UNIFORMLY MOVING CHARGE VIA L ORENTZ TRANSFORMATION In the relativistic formulation of classical electrodynamics the E and B eld vector form the antisymmetic electrodynamic eld tensor 0 Ex Ey Ez Ex 0 cBz cBy Ey cBz 0 cBx Ez cBy cBx 0
Show that the elds from a charge q in uniform, rectilinear motion q 1 v 2 c 2 R0 E 40 s3
are obtained via a Lorentz transformation of the corresponding elds in the rest system of the charge.
Solution
We wish to transform the EM elds. The EM elds in a covariant formulation of electrodynamics is given by the electromagnetic eld tensor
where we are using components running as 1, 2, 3, 4. To transform the EM elds is to transform the eld tensor. A Lorentz transformation of the eld tensor can be written
where
hjj i
0 0
0 1 0 0
0 0 1 0
0 0
mm l
~F0 L L
L L F
! hjjk
0 cB3 cB2 E1
cB3 0 cB1 E2
cB2 cB1 0 E3
E1 E2 E3 0
mm l
E c2
&
! %
A A
(3.31)
(3.32)
(3.33) (3.34)
(3.35)
3
E XAMPLE 3.4
mm l
! i hkj ! j
0
E ND
(3.30)
OF EXAMPLE
3.3
24
so that F 0 0 0 0 E1 0 0 0 0 E2 0 0 0 0 E3
0 E1 0 E2 0 E3 0
A little matrix algebra gives 0 0 0 0 E1 0 1 0 0 0 0 0 0 E2
where s
vt
&
"
ct
hjjk i n
hjjk
y z
&
! %
n
x0 y0 z0 ct 0
x y z ct
mm sl
mm sl
mm sl
!
B3
B2
B1
E1 E2 E3
0 E1 0 E2 0 E3
0 0 E c 3 0 E c 2
&
"
! %
0 0 2 E1 2 E1 0 E2 0 E3 2 0 1 2 E1
0 E2 0 0 0 E2
0 E3 0 0 0 E3
0 2 E1 0 E2 0 E3 2 0 0 E1 2 E1
2 1
n
mm l
! hkj ! j
0 0
0 0 1 0
0 0
0 E1 0 E2 0 E3 0 E1
0 0 0 0 E2
0 0 0 0 E3
0 E1 0 E2 0 E3 0 E1
mm l
n
jhj
mm l
&
"
! %
jhj
hjj t
hjjk t
~FL L
~ L
0 0 0 0 E3
0 E1 0 E2 0 E3 0
0 0
0 1 0 0
0 0 1 0
0 0
mm l
mm l
mm sl
hjjk i
B0
"
"
E0
"
q x0 x y0 y 40 x0 2 y0 2
z0 z z0 2
3
"
hjjk
where
2 and
v c where v
v1 x. The elds in the rest system S0 are (3.36) (3.37)
(3.38)
(3.39)
(3.40)
(3.41)
(3.42) (3.43) (3.44) (3.45) (3.46) (3.47)
(3.48)
3.3. S OLVED
EXAMPLES
25
E R2 0
q x s x y y z z 40 2 x s 2 y2 z2 3 s x
2
y2
z2
&
! %
q 1 2 40 R3 0
R0 1
2 sin2
&
! %
R2 1 0
2 sin2
w&
%&
"
% "
R2 0
y2
z2
1 2
"
" &
! %
R2 0
y2
z2
"
"
y2
z2
" &
! %
y2 z2 2
q 40 2
6&
! %
s x
R0 sin
R0 cos
3
R0
&
"
"
" &
"
&
" " & ! % ! % % ! u%
(3.49) (3.50) (3.51) (3.52)
y2
z2
(3.53) (3.54) (3.55) (3.56)

E ND 3.4
OF EXAMPLE
26
L ESSON 4
Lagrangian and Hamiltonian Electrodynamics

4.1 Coverage
We briey touch the Lagrangian formulation of electrodynamics. We look at both the point Lagrangian for charges in EM elds and the Lagrangian density of the EM elds.
4.2
Formulae used
q
The Lagrangian for a charged particle in EM elds is
A useful Lagrangian density for EM eld and its interaction with charged particles is given by
z
E2
A j
1 c2 B2 2 0

qv A
mc2
(4.1)
(4.2)
27
28
L ESSON 4. L AGRANGIAN
AND
H AMILTONIAN E LECTRODYNAMICS
4.3
E XAMPLE 4.1
Solved examples
C ANONICAL QUANTITIES FOR A PARTICLE IN AN EM FIELD Derive the canonical momentum and the generalised force for the case of a charged particle mc2 in EM eld given by and A. The Lagrangian is L q qv A.
Solution
We know from analytical mechanics that the canonical momentum P is found through L Pi (4.3) vi so with mc2 q qv A (4.4) L we nd that
m vi qAi P p qA
qv A, which gives us,
"
qE
qv
"
"
qv
A t
&
! &
"
&
~
"
qv
qv
qv

#~
! &
}
"
!
v A
"
xi
v jA j xi
dAi dt dA q dt
" |
!
Qi
q xi
qv j A j
d q dt vi
qv j A j
A t
And on the other hand the generalised force is U d U Qi xi dt xi where U is the generalised velocity dependent potential q
c2
vjvj c2
"
"
mc2
vi
qAi (4.5) (4.6)
"
"
"
'
!
!
!
!
Pi
L vi vi
mc2
vi vi c2
qv j A j
"
!
"
!
(4.7)
(4.8)
(4.9)
4.3. S OLVED
EXAMPLES
29
What does the canonical momentum P p qA represent physically? Consider a charge moving in a static magnetic eld. This charge will perform gyro-harmonic motion. Obviously the momentum is not conserved but on the other hand we did not expect it to be conserved since there is a force on the charge. However we now from analytical mechanics, that it is the conservation of canonical momentum that is more general. Conservation of canonical momentum is found when the problem is translational invariant, which is true in the case we have here since the potentials do not depend on spatial coordinates. So we expect P p qA to be a constant of the motion, but what is it?
G AUGE
INVARIANCE OF THE
L AGRANGIAN
DENSITY
Consider the Lagrangian density of the EM elds in the form
We know that EM eld are invariant under gauge transformations
Determine the Lagrangian density under a gauge transformation. Is it invariant? If not, discuss the consequences this would have.
Solution
Let us insert the gauge transformation relations into the Lagrangian density. Remembering that E and B are invariant under gauge transformations we nd that E2 A
! &
(4.17)
! &
! &
E2
A j
"
! &
E2
! &
1 c2 B2 2 0 1 c2 B2 2 0 1 c2 B2 2 0
E2
A j A j
% % ! & " " T " T $ % ! & "& % ! T " T $ % ! T ! & " T " T % ! &
&
! T %
"
&
$
! T
! T
" T
1 c2 B2 2 0
"
! &
E2
A j
1 c2 B2 2 0
(4.10)
(4.11) (4.12)
(4.13) (4.14)
(4.15) t (4.16) t
"
"
!
E ND
OF EXAMPLE
4.1
E XAMPLE 4.2
30
L ESSON 4. L AGRANGIAN
AND
H AMILTONIAN E LECTRODYNAMICS
where we have used the continuity equation. We see that as , but what about the other two terms?
Some thought reveals that the neccessary condition for a Lagrangian to be physically acceptable is not the Lagrangian itself is invariant but rather that the variation of the action integral S d3x dt is invariant. So now we would like to check that the gauge transformations indeed do not affect any the variation of the action. Now it is possible to proceed in two different ways to do this: one is simply carry out the integration in the denition of the action integral and check that its variation is zero, or two, remembering that the variation of the action is equivalent to the Euler-Lagrange equations, one could plug in the Lagrangian density (4.17) into Euler-Lagrange equations to check the resulting equations differ from the Maxwell equations. Let us use the rst alternative. Since the action is linear in
where we have used the continuity equation. Furthermore, if we assume no ux source/sink at innity then we can write
Now when taking the variation, we realize that we must hold t0 and t1 the end point of the particle path xed, and thus
As one would expect, gauge transformations of the potentials do not effect the physics of the problem!
Sext
F
Sext
d3x
t1
t0
F "
j dS dt
d3x
t0 t1 t0
"
F F !
F $"
j dS dt
d3x

'
" @&
F F
Sext
d3x dt
t1
obviously has the same form
$V'
F F
F F
#
it is sufcient to examine
d3x dt (4.18)
(4.19)
(4.20)
E ND
OF EXAMPLE
4.2
L ESSON 5
Electromagnetic Energy, Momentum and Stress

5.1 Coverage
Here we study the force, energy, momentum and stress in an electromagnetic eld.
5.2
Formulae used
E
Poyntings vector
Maxwells stress tensor
Ti j
Ei D j
(5.1)
Hi B j
1 ED 2 ij k k
Hk Bk
(5.2)
31
32
L ESSON 5. E LECTROMAGNETIC E NERGY, M OMENTUM
AND
S TRESS
Table 5.1. The following table gives the relevant quantities. The eld vectors in this table are assumed to be real. If the given elds are complex, use the real part in the formulas.
1 2E
5.3
E XAMPLE 5.1
Solved examples
QUANTITIES POTPOURRI
EM
Determine the instantaneous values of the energy density, momentum density, intensity and stress associated with the electromagnetic elds for the following cases:
Solution
Background (a) Progressive wave This case is an example of a progressive or propagating wave.
Since E and H are complex we must rst take their real parts:
The energy density is
&
!
Re H
k1 E cos k1 x1 0
&
Re E
E0 cos k1 x1
t e2 t e3
Also, identify these cases. Assume that D
c)
Br B
0 m cos 2 r 3 0 m sin 4 r 3
b) same as in a) but for k1
E and B
a)
for k1
E H
E0 e2 ei k1 x1 k E k
, k2
k3
0.
H.
&
"
"
Ei D j
"
Name Energy density Intensity Momentum density Stress
Symbol Uv S PEM T
Formula D 1H B 2 E H E H Hi B j 1 i j Ek Dk 2
Hk Bk
SI unit J/m3 W/m2 kg/s m2 Pa
(5.3) (5.4)
5.3. S OLVED
EXAMPLES
33
The intensity or power density is
t e1
The momentum is
The stress is
(b) Evanescent wave This case is a example of an evanescent wave. We take real part of the elds keeping in mind the fact that k1 i is imaginary:
The energy density is
! 6 &
!
Re H
E0 e
!
Re E
E0 e
x1
sin t e2
e Im E 1 E0 x1 e cos t e3
!
T33
T32
e 1
x1
&
! &
T22
T21
&
! &
&
!
T11
2 2 E H 2 2 2 3 2 2 k1 2 E cos2 k1 x1 t E0 cos2 k1 x1 t 2 2 0 2 E0 cos2 k1 x1 t T31 T12 0 2 2 E H 2 2 2 3 2 2 k1 2 E0 cos2 k1 x1 t E cos2 k1 x1 t 2 2 0 T13 T23 0 2 2 H E T22 0 2 3 2 2
&
!
!
PEM
2 E0 cos2 x1 t
t e2
&
&
2 E cos2 x1 0 2 E cos2 x1 0
e3
cos t e2
6&
&
!%
&
&
%
E0 cos k1 x1
t e2
k1 E cos k1 x1 0
t e3
w&
&
" &
!
!

Uv
2 2 k1 E 2 cos2 k1 x1 E0 cos2 k1 x1 t 2 2 2 0 2 E0 cos2 x1 t
(5.5) (5.6)
(5.7) (5.8) (5.9)
(5.10)
(5.11) (5.12)
(5.13) (5.14) (5.15)
(5.16)
(5.17)
34
AND
S TRESS
The intensity is
The momentum is
The stress is T11
(c) Magnetic dipole This case is a magnetic dipole. The elds are real and in spherical
coordinates. The energy density is
2 B
0 m2 1 4 cos2 sin2 6 32 2 r 0 m2 1 3 cos2 32 2 r6
The intensity is S components are
0 since E
0, and likewise for the momentum. The stress tensor
"
&
"
6&
" %
"
Uv
'
1 B2 2 0 r
1 2 0
2 0 m2 4 cos2 16 2 r6
2 0 m2 sin2 16 2 r6
!
!
T33
T22
'
2 H 2 3
2 E 2 2
T32
2 E0 e 2 x1
sin2 t
cos2 t 2
2
T22
'
2 2 E H 2 2 2 3 T13 T23 0
T21
2 E0 e 2 x1
sin2 t
cos2 t 2
2
"
'
2 E0 e
2 2 E H 2 2 2 3 T31 T12 0
!
!
PEM
!
2 E0 e
2 x1
sin t cos te1
2 E0 e
2 x1
sin t cos te1
2 x1
sin2 t
2 cos2 t 2
"
"
'

Uv
2 2 2 x1 2 2 E0 2 x1 e sin t cos2 t E0 e 2 2 2 2 E0 e 2 x1 2 sin2 t cos2 t 2 2
(5.18) (5.19)
(5.20)
(5.21)
(5.22) (5.23) (5.24) (5.25) (5.26)
(5.27)
5.3. S OLVED
EXAMPLES
35
0 m2 cos2 4 2 r6 0 m2 1 32 2 r6
C LASSICAL ELECTRON RADIUS Calculate the classical radius of the electron by assuming that the mass of an electron is the mass of its electric eld and that the electron is a homogenous spherical charge e. Mass and energy are related through the distribution of radius re and total charge q equation E mc2 .
Solution
One the failures of Maxwells equations or classical electrodynamics is on question of mass of point particles. From relativity one can show that fundamental particles should be point-like. However, if one calculates the electromagnetic mass from Maxwells equation one gets an innite result due to the singularity in Gauss law E 0 . This points to the fact that Maxwells equations has a minimum length scale validity where quantum mechanics takes over. We will calculate this problem as follows: determine the electric eld in all of space and then integrate the formula for the energy density of the electric eld. This integral will contain the radius of the electron since it partitions the integration. We then relate this eld energy to the mass of the electron, which is a known quantity. We use this relation to solve

Tr T T
&
0 m2 2 sin2 1 3 cos2 32 2 r6 0 m2 1 3 cos2 2 sin2 32 2 r6 0 0 0 0
&
" %
&
"
! %
!
T r Tr
1 2 B 0
Uv
T r
0 m2 sin2 16 2 r6
0 m2 1 32 2 r6
&
(5.28) (5.29) (5.30) (5.31) 3 cos2 (5.32) (5.33) (5.34) (5.35) (5.36)
E ND
OF EXAMPLE
0 m2 8 cos2 1 3 cos2 32 2 r6 0 m2 1 5 cos2 32 2 r6 0 m2 sin cos 1 Br B 0 8 r6 0 T r
&
5.1
" %
&
! %
!

Trr
1 2 B 0 r
Uv
3 cos2
E XAMPLE 5.2
36
AND
S TRESS
for the electron radius. We start by determining the electric eld. Consider a homogeneous spherical charge distribution in a vacuum. Introduce spherical coordinates. We divide the whole space into two parts: the region outside of the sphere, i.e. r re , is given by Coulombs law: e E r re (5.37) 40 r2 and the region inside the sphere is given by Gauss law which in this case gives
If we integrate this equation, i.e. (5.39), we nd that e r Er r re 3 4 re 0
We can easily verify that (5.40) is indeed a solution to (5.39) and furthermore it is continuous with (5.37), thus making the solution unique. We may now determine the energy density of the electric eld due to the electron. It is simply
Uv r
re
Now we integrate Uv over all space. That is all space

4 0 re 0
re
Finally, we relate the total electric eld energy to the rest mass of the electron and solve for the electron radius,
re
3 e2 5 40 me c2
me c2 3e2 1 200 re
me c 2 (5.45)
"
e2 r5 6 80 re 5
re
e2 80
1 r
3e2 1 200 re
F "
e2 r2 r2 dr d 6 32 2 0 re
&
R %
" &
I %
F
Uv r
6& 6&
I % R %
F
Uv
1 E2 2 0 Uv r
|
1 2 r Er r2 r
e 3 4 3 r e 0
re
6&
6&
I %
Er
re
1 e 32 2 0 r4 2 e 2 6 r 32 2 0 re
re
Uv r
6&
I %
R %
U 9 9
(5.38)
(5.39)
(5.40)
(5.41) (5.42)
re d3x
re
e2 1 2 r dr d 32 2 0 r4 (5.43)
(5.44)
5.3. S OLVED
EXAMPLES
37
This last result can be compared with the de facto classical electron radius which is dened as
and is found by calculating the scattering cross section of the electron.
S OLAR SAILING Investigate the feasibility of sailing in our solar system using the solar wind. One technical proposal uses kapton, 2 mm in thickness, with a 0.1 mm thick aluminium coating as material for a sail. It would weigh 1 g/m2 . At 1 AU (= astronomical unit distance between sun and earth) the intensity of the electromagnetic radiation from the sun is of the order 1 4 103 W/m2 . If we assume the sail to be a perfect reector, what would the acceleration be for different incidence angles of the suns EM radiation on the sail?
Solution
In this problem we will use the principle of momentum conservation. If the continuity equation for electromagnetic momentum PEM t T Pmech t is integrated over all space the stress term disappears and what is left says that a change in electromagnetic momentum is balanced by mechanical force. Imagine a localised pulse of a progressive electromagnetic wave incident on a plane. That the wave is progressive means simply that it is purely radiation (more about progressive waves in the next lesson). Let us characterize this wave by a Poynting vector S of duration t. It travels in space with velocity c, it lights up the area A on the surface on the sail, and, furthermore, S makes an angle 180 with the normal of the sail surface. The momentum carried by such a wave is
so the momentum along the direction of the surface normal n is
After hitting the sail, the pulse will be characterized with all the same quantities as before the impact, except that the component along the surface normal will be opposite in sign. This is because the sail is assumed to be perfectly reecting. Thus,
Now the continuity equation for EM momentum says that PEM t the component along the sail surface normal this implies that
!
pafter ^ EM n
pbefore ^ n EM
!
pbefore n EM
S A t cos2 c
S A t cos2 c
&
6&
pbefore EM
Pbefore V EM
S c2 A c t cos
(5.47)
(5.48)
(5.49)
Pmech
F so for

"
re
e2 40 me c2
(5.46)
E ND
OF EXAMPLE
5.2
E XAMPLE 5.3
38
AND
S TRESS
pbefore S EM 2 n ^ 2 c A cos t The other force components are zero by symmetry. n F ^ A
We now have the pressure exerted by a pulse charcterized by S incident on a surface with an angle . The solar wind can be seen as a multitude of such pulses radiating radially outwards from the sun. At 1 AU, the solar constant is 1.3 kW m2 . The solar constant is the intensity of EM radiation or in other words, the magnitude of the Poynting vector S . From equation (5.51), we nd that
Newtons second law gives the acceleration a of the sail (which, as we recall, weighs 3 10 3 kg/m2 ) due to the solar wind can at most be
which is half of the acceleration due to the suns gravitational eld.
0 9 10 5 N/m2 3 10 3 kg/m2
13 3
03 W/m2 cos2 108 m/s
(5.52)
3 mm/s2
(5.53)
E ND
OF EXAMPLE
5.3
So nally, the pressure
exerted by the pulse on the sail is simply S cos2 c (5.51)
F ^ n
|
pafter EM
(5.50)
b
5.3. S OLVED
EXAMPLES
39
M AGNETIC PRESSURE ON THE EARTH Determine the magnetic pressure due to the Earths magnetic eld at the magnetic poles (take B 6 10 5 T) and compare this with the Earths atmospheric pressure (1 atm 1 01 105 Pa). Now assume that the magnetic dipole moment is proportional to the angular velocity of the earth, how much faster would the angular velocity need to be for the magnetic pressure to be comparable to the atmospheric pressure?
Solution
In this exercise we see that EMF can exert pressure not only via radiation pressure, but also through static elds. The static magnetic pressure is quantied in the Maxwellian stress tensor, which for this case is
So by taking the inner product of T and the unit vector pointing in the direction the north pole, namely 1 0 0 , we nd the pressure in the radial direction to be
Now we assume, according to the hypothesis in problem, that
T1
24 3600
1 4 10 105
So with
105 Pa, we nd that

3
T1
T0
0 1
10 s
E ND
OF EXAMPLE
where we denote the current values of the pressure and rotational period time, respectively and the hypothetical values 1 and T1 . Solving for T1 we arrive at
U
so that
P0 P1
0
1 mB Tperiod
1 4 10
B2 r 2 0
6 10 5 2 2 1 3 10 1 3 10 6
b A % & b
t
B2 0 r 0 0
0 0 0 0 0 0

E XAMPLE 5.4
Uv 0 0
0 0 Uv 0 0 Uv
(5.54)
Pa
(5.55)
(5.56)
(5.57) and T0
(5.58)
(5.59)
5.4
40
L ESSON 6
Radiation from Extended Sources

6.1 Coverage
We will study the important general problem of how to calculate the EM elds induced by spatially extended, time-varying sources. This problem is solved in different ways depending on the explicit form of the source distribution. For truly extended bodies with non-monochromatic time dependence, we use the general expressions for the retarded potentials. And for monochromatic, one-dimensional current distributions, which we will call antennas here, we use the formulas given below.
6.2
Formulae used
At x d3x
General expressions for retarded potentials
For antennas, we use
ik sin eik x x 40 c x x
i0 k sin eik x x 4 x x
41
z y
t x
1 40
d3x
y z
x t x x
y z
jx t x x
0 4
z y
42 k2 sin2 0 c 32 2 x x 2
L ESSON 6. R ADIATION
FROM
E XTENDED S OURCES
where
l l
6.3
Solved examples
E XAMPLE 6.1
I NSTANTANEOUS
CURRENT IN AN INFINITELY LONG CONDUCTOR
Consider the following idealised situation with an innitely long, thin, conducting wire 0, it is current free, but at time t 0 a constant current J is along the z axis. For t applied simultaneously over the entire length of the wire. Consequently, the wire carries the current 0 t 0 j z t 0 J
Hint: Calculate rst the vector potential A .
Solution
This problem belongs to the most general category of problems of the kind where given a source distribution one wants to nd the EM elds. This is because the source is not monochromatic, so it is not an antenna, and furthermore it is an extended distribution, so multipole expansion analysis is not possible. So we must use the most general formula for calculating elds induced by time-varying sources, which in the Lorentz gauge take the form
6& P %
t x
d3x
T ! T &T PT %
1 40
a& P %
At x
d3x
T ! T & hT P T %
0 4
jt x x x
t x x x
It is assumed that the conductor can be kept uncharged, i.e., in whole space.
6& %
j z e
ikz cos
y z
dz
0 . Determine B, E and S
(6.1)
(6.2)
6.3. S OLVED
EXAMPLES
43
where the source time t is to be replaced by t t x x c so t is seen as a function of t, x and x . The function t x x c is known as the retarded time tret , an expression which is meaningful only relative the eld point t x . The algorithm of solution now that we have decided to use (6.1) is to nd the explicit form for j t x and t x , perform the integrations to obtain the potentials A t x and t x and, nally, derive the E and B elds from the potentials in the normal fashion and the Poynting vector from the elds. Let us nd the explicit expression for the current density j x t , which is illustrated in Fig. 6.1. In many problems, the expressions for the sources consist of a time-dependent part times a space-dependent part. This is one such case. The switching on at t 0 can be written as a step function t . And if we orient the wire along the z axis, the space dependent part can be written x y J z. On the other hand, the charge distribution t x 0 as given in the problem formulation. This can be seen as and charges owing in opposite directions such as to keep the total charge density t x 0 but this could still have a total current density j t x 0. So we have for the current that and for the charge density
(6.5)
Figure 6.1. The current density distribution j is along the z axis and is turned on at t 0. We use cylindrical coordinates. For the remaining z integration we use cylindrical coordinates (see Figure BLP1.2cyl) so we can write
T
! !
We insert (6.3) into (6.1), remembering to replace the source time t with t and perform the integration. The integrations over x and y are trivial: J t x zz c z 0 At x dz 4 x zz
T ! & T ! ! %
T
b F
6& T P T %
t x
0 x
&T %
&T % &T %
6& T P T %
jt x
x y J t z
S P % &
& P %
T
& hT P T %
T
! !
& P %
6& P %
qT
& hT % hT % & & T %
T
! !
& hT P T %
6& P %
"
& hT P T %
6& P %
& P %
(6.3) (6.4) x c,
44
FROM
E XTENDED S OURCES
The step function in the integrand is zero when its argument is less than zero. This means that when integrating over z , only those z contribute which satisfy which can be written, if we assume as
or where we have introduced a simply as a shorthand. These limits can be understood as follows (cf. Fig. 6.2): after the current is switched on, EM elds are sent from each point along the wire and travel at the speed of light. Since the information (i.e., the current turnon) carried by the elds travel in the line of sight, or in other words in a straight line, each eld point only sees those parts of the current which are close enough. This illustrates the concept of retarded time, which is only meaningful relative the eld point, and also the information gathering sphere.
P r
Figure 6.2. This series of snapshots shows what the part of the current is seen at the eld point P at different eld times t. So we have now that J a A z 0 4 a
z2
r2
J z 0 ln 4
r2 c2 t 2 r2 c2 t 2
dz
r c
r2 c
I T
c2 t 2
r2
I T I " T
z2
r2
c2 t 2 r2
c2 t 2
c2 t 2
r2
"
T
rr
zz c
ct
r2
A T
z2
"
T
x x
rr x
r2
z2
(6.6) (6.7)
9 9 9 !
(6.8) (6.9)
(6.10) (6.11)
(6.12)
r2 2
(6.13)
6.3. S OLVED
EXAMPLES
45
Now that we have the vector potential we can derive the E and B elds. The B eld is
and since A only has its z -component different from zero in the cylindrical system, we have A z (6.15) B r
we then can write
The electric eld is derived from
But
and so
2
and
All we have left is to determine the Poynting vector S
Notice that as t
then E
!
0 J 2 t
1 1
r2 c2 t 2
0!
&
! %
"
) 7)
ln
"
'
'
1 1
f f
f t
1 2r2 2 f c2 t 3
1 r2 f c2 t 3 1 1 f 2 t f 1 f
f
) )
ln
"
'
'
!
!
A t J z 0 4 t
1 1
f f
2 ft
&
%
Observe that for large t we have the static case: B
2 r
0 J
1
r2 c2 t 2
) )
"
'
'
!
Az r
!
f r
c2 t 2
r 1 f
0 J ln 4 r 0 J 2 4 r f
1 1
f f
0 J 2 r
&
%
Thus we only need
Az r .
Introducing f
!
r2 c2t 2 , which satises (6.16)
& 6hT P T %
6& P %
On the other hand, the scalar potential we may set so t x
0, since t x
0.

(6.14)
(6.17) (6.18)
(6.19)
(6.20)
(6.21)
(6.22)
46
L ESSON 6. R ADIATION 1 0 J 0 2 r 1
FROM
E XTENDED S OURCES
0 J2 r 2 2 rt 1 cr22 2t
From this expression we can for example calculate the radiated power per unit length by integrating over a cylindrical surface C enclosing the wire:
We see that an innite power is transmitted starting at t 0 and r 0 which travels out to innite r. So in practice it is impossible to produce this physical setup. This is due to the quick turn on. In the physical world only gradual turn ons are possible. This many be seen as a consequence of what is known as Gibbs phenomenom. In the above we tacitly employed the retarded potential without discussing the possibility of using the advanced potential. Let us see what happens if we apply the advanced potential to this problem. The only thing that changes from the outset is that the source time t is replaced by the advanced time ta t x x c instead of the retarded time tr t x x c. With ta as the argument to the step function, the contribution to the integral comes only from those z which satisfy
If we assume that
then we may write this as
or
One can proceed further and calculate the resulting integral. But what is interesting is that now we see that the relation (6.9) seems to say that we have no information about what happened before turn on, while the relation (6.27) says we have no information about what happened after turn on. Physics seems to be conspiring on us in such a way that we cannot compare the advanced and the retarded potential at the same time!
! 0I T
c2 t 2
r2
R T " T
z2 z
r2
c2 t 2 c2 t 2 r 2
c2 t 2
r2
T !
T !
"
rr
zz c
rr
zz c
T
! U"
&
! %
T
(
S dC
0 J2 2 t 1 cr22 2t
r2 c2 t 2
2 t
r2 c2 t 2
&
E ND
&
! %
&
! !

1 E 0
0 J
(6.23)
(6.24)
(6.25)
(6.26)
(6.27)
(6.28) (6.29)
(6.30)
OF EXAMPLE
6.1
6.3. S OLVED
EXAMPLES
47
M ULTIPLE HALF - WAVE ANTENNA half wavelengths is assumed to have a current disA wire antenna with a length of m tribution in the form of a standing wave with current nodes at the its endpoints. Determine the angular distribution of the radiated electromagnetic power from the antenna. Hint: It can me convenient to treat even m and odd m separately.
Solution
One realises that the setup in this problem is an antenna, since we have a monochromatic source and the current is an extended one-dimensional distribution. Thus, we may use the antenna formulae, but since only the radiated effect is ask for all we need is
where
l
so we seek a form for the current distribution j z . It simplies matters if we consider the cases for m an even and m an odd multiple of half wavelengths separately. As shown in Figure 6.3, for the case of m even, we will use sin kz and for m odd we use cos kz .
Figure 6.3. Depending on whether the length of the wire is an even (as seen in a) or odd (as seen in b) multiple of half the wave length, the current distribution is sin kz or cos kz . We perform both integrations over z from l to l which is the total length L, which is a multiple m of 2. These facts give us a relation between the integration limits and k, namely m L 2l m 2 l (6.33) 2 k
& T %
& T %
& T %
& T %
a) even
sin kz
b) odd
cos kz
& T
&T %
j z e
& T %
& % b T
ikz cos
b 7# F
6
k2 sin2 0 c 32 2 r2
dz
& %
E XAMPLE 6.2
(6.31)
(6.32)
48
FROM
E XTENDED S OURCES
Let us consider the odd case. The current distribution is written
and so
m 2
m 2
m m sin 1 cos sin 1 cos 2 2 cos sin 1 cos sin 1 cos 2 2 J0 2 sin m 2 cos m 2 cos 2 cos cos m 2 sin m 2 cos k sin2 2J0 cos m 2 cos (6.35) k sin2
Inserting this into (6.32) gives us

2 J0 0 c cos2 m 2 cos r 8 2 r2 sin2
Let us consider the case when m is even so the current distribution is written
We remember the relation (6.33) which is still valid but now m is an even number. So we have that
AT
6& T %
j z
(6.36)
J0 sin kz
(6.37)
&
&
&
Q &
" %
" @&
" %
" &
! &
&
&
&
! %
! %
J0 k sin2
b &
! % b d b
&
J0 eiu 1 cos 2k i 1 cos
e iu 1 cos i 1 cos
"
eiu
iu
! % b d { b F
J0 2k
m 2
iu cos
du
m 2
T
dz
T
J0 cos kz e
dz
b T
T
ikz cos
b (v F
6& T % "
j z
J0 cos kz
(6.34)
kz 1 k du
6.3. S OLVED
EXAMPLES
49
m 2k
m 2
2 J0 0 c sin2 m cos 2 r 8 2 r2 sin2
( 9
2 4J0 m sin2 cos 2 sin4 2 k
(6.39) (6.40)
E ND
OF EXAMPLE
6.2
"
2J0 m m cos sin cos 2 2 2 ik sin 2J0 m cos i sin 2 2 k sin
&
" %w
&
! %w
sin
cos sin
cos
m 2
m 2 m m cos sin cos cos 2 2 cos
&
" y " %w
&
sin
sin
! y
cos
m 2
cos
m 2
(6.38)
&
" %w
&
&
! %w
&
! %w " %
! % ! y
"
J0 ik sin2 J0 ik sin2
cos sin
m cos 2
cos sin
cos
" %
! %
i1
i1
J0 ei 1 2ik
e i1 cos
ei 1
e i1 cos
m 2
& v d b
v d
& b d b
cos
m 2
cos
m 2
! e
eiu 1
v d b
b d
m 2
cos
m 2
iu 1 cos
eiu
b d
b F
b F
J0 2ik J0 2ik
m 2
iu cos
iu
du du
cos
m 2
cos
m 2
T
J0
sin kz dz
T
m 2k
ikz cos
u dz
kz 1 u du
!
v 9
m 2
m 2
50
FROM
E XTENDED S OURCES
E XAMPLE 6.3
T RAVELLING WAVE ANTENNA A wire antenna of length is fed at one of its endpoints by a transmitter signal and is at its other end terminated with a resistance to ground. The termination is adjusted such that no current is reected back on the wire. This means that the current distribution comprises travelling waves emanating from the feed point so one can assume that j z J exp ikz along the wire. Determine the angular distribution of the electromagnetic radiation from this antenna.
Solution
We need the formula k2 sin2 32 2 r2
L 2
where dz e
L 2
L 2
Finally we have
&
! {%
|&
! %
2 J0 sin2 8 2 r2
0 sin2 kL 1 cos 2 0 1 cos 2
)&
! %
&
! %
(
sin2
'
2 4J0 k2 1 cos
kL 1 2
cos
) &
! %
2 J0 sin k 1 cos
kL 1 2
cos
L 2
' b &
&
! %
J0
eik 1 cos z ik 1 cos
g e
! % b d
b F
J0
L 2
g e
dz eik 1
b d
L 2
cos z
& wT %
In this case the distribution is given: j z we get
sin
&T % g
ik cos z
b T
b F 7#
0 0
(6.41)
j z
(6.42) J0 eikz . If we insert this distribution into (6.41),
1 i e 2i
(6.43)
(6.44)
(6.45)
E ND
OF EXAMPLE
6.3
& hT
& whT %
v
6.3. S OLVED
EXAMPLES
51
Microwave links are based on direct waves, i.e., propagation along the line of sight between the transmitter and receiver antennas. Reections from the ground or a water surface may in unfavourable cases distort the transmission. Study this phenomenon using the following simple model: The transmitter antenna T is a horizontal half-wave dipole placed a distance h1 above the ground level. The receiving antenna R is in the main lobe from T, at a horizontal distance D from T, and at height h2 . The signal at R is considered to be composed of the direct wave and a wave reected from the ground. The reection is assumed to cause a phase shift in the wave, but no loss of power. The ground is considered at over the distance D, and h1 h2 D. 1. Calculate the electric eld E (magnitude and direction) in at R if the transmitter antenna is fed with a current j. 2. Discuss the meaning of the result.
T h1 h1
Solution
The Fourier transform of the E eld in the far zone (radiation eld) from a half-wave dipole antenna, i.e., a linearly extended current distribution with length 2 in the z direction,
4
Since k k r and we study the radiation in the maximum direction, i.e., perpendicular to j z J cos kz z so that k x 0, this expressions simplies to
r1
r2
b F
fT
a& %
E rad r
dz eik x j
g c
i eikr r 40 c r
3. Simplify the result for the case h1 h2
& T %
& wT %
M ICROWAVE
LINK DESIGN
E XAMPLE 6.4
D.
h2
52
FROM
E XTENDED S OURCES
A superposition of the direct and reected contributions (with the distance from the transmitter T to the receiver R equal to r1 and r2 , respectively), with due regard to the phase shift (corresponding to a change of sign in the current), gives the Fourier transform of the total E eld at the far zone point R :
2D 2D
where
and 2h1 h2 (6.48) D is the difference in path distance. Insertion of (6.47) and (6.48) into (6.46), with k c 2 , gives the Fourier transform of the eld at R
from which we obtain the physical E eld 2D
We see that the received signal at R will be extinct if
2 h1 h2 D
E ND
OF EXAMPLE
! )
"
6& P %
Etot r t rad
tot Re E rad e
"
'
i t
J 2 h1 h2 2 sin cos 0 cD D
6& %
Etotrad r
"
2D
iJ eikr e 20 c D
"
h2 1
h2 2
ikr 2
eikr
2 h1 h2 J eikr sin z 0 c D D h2 1 h2 2
"
"
"
"
% "
r2
h1
h2
&
D2
"
&
"
h1
h2 2D
h2 1
h2 2
h1 h2 D
"
% "
% "
r1
h1
h2
&
D2
"
&
% "
Since h1 h2
D, Pythagoras theorem gives h1 h2 2D

2
h2 1
h2 2
6& %
" & %
6& %
Etotrad r
Edirrad r
Ererad r
'
iJ 20 c
t b & sq % r T
ik eikr J sin kz 40 c r k 2
4 4
iJ eikr z 20 c r
eikr1 r1
eikr2 r2
&
z (6.46) h1 h2 D r r 2 r 2 (6.47)
s& T %
b F p
Q& %
E rad r
dz J cos kz r
ik eikr 40 c r
t z
6.4
L ESSON 7
Multipole Radiation
7.1 Coverage
We look at electric dipole, magnetic dipole and electric quadrupole radiation. Multipole radiation analysis is important since it simplies the calculation of radiation elds from time-varying eld and since EM multipoles exist in many elds of physics such as astrophysics, plasma physics, atomic physics, and nuclear physics.
7.2
Formulae used

0 eik x p 4 x 1
Fields in far regions from an electric dipole B rad x
Fields in the far zone from a magnetic dipole B rad x k
E rad x
Field in the far zone from an electric quadrupole

k eik x m 40 c x
0 eik x m 4 x
E rad x
p 1
1 eik x 40 x
z y z y z y z y
53
54 i0 eik x 8 x
L ESSON 7. M ULTIPOLE R ADIATION
7.3
E XAMPLE 7.1
Solved examples
R OTATING E LECTRIC D IPOLE An electric dipole with constant electric dipole moment magnitude is located at a point in the xy plane and rotates with constant angular frequency.
(a) Determine the time-dependent electromagnetic elds at large distances from the
dipole.
(b) Determine the radiated average power angular distribution and the total radiated
power.
Solution
(a) We can write the time-varying dipole momentum relative the location of the dipole
as
which represents a constant dipole moment p0 times unit vector rotating with angular frequency . This can also be rewritten in complex form or
where c.c., stands for the complex conjugate of the term opposite the sign. In what follows we use the convention that we write the dipole expression as a complex quantity but we drop the c.c. term, which is commonplace when discussing harmonic oscillation. It is easy to identify the Fourier component of the dipole moment in this case
We notice, in this complex space variable space, that the y component has the phase factor i ei 2 relative the x component, which is due the circular rotation. We would like to express (7.4) in spherical components rather than Cartesian components since the expressions for dipole elds in spherical components are simpler, so we transform
"
6& %
p t
p0 x
ip0 y
"
A A
" &
"
% 6& %
pt
p0 e
x ip0 e
i t
"
6& %
pt
Re p0 e
ip0 e
i t
cc
i t
&
"
6& %
pt
p0 cos t x
sin t y
i t
E rad
k Q
y
B rad
k Q
i eik x 80 x
(7.1)
(7.2)
(7.3)
(7.4)
7.3. S OLVED
EXAMPLES
55
the base vectors x and y in the conventional fashion and get
Now that we have the Fourier component of the dipole moment expressed in spherical components we insert this into the dipole radiation fomulae:
First we calculate
from which we get p k
so nally we can write the eld in space and time coordinates (remember: Bt x Re B x e i t ),
So our nal expression is
& &
"
! &
"
sin kr
& &
"
" &
"
0 2 p0 sin kr t 4 cr 2 p0 cos cos kr t 40 c2 r
cos cos kr
}&
&
e v
b d
! 1& P %
Bt x
0 ei kr t Re p0 kei i cos 4 r 0 p0 k Re ei kr t ei i cos 4 r
&
p0 k2 ei
cos
b d
! %
% b &
p0 ke
r 0 i 1 0
&
p0 kei
! %
!
&
p0 ke
cos 0
r sin cos 1 0 i cos
&
!
!
0 eikr p k 4 r 1 eikr p k 40 r
"
p0 e
i 0
"
"
i cos r sin cos
sin
cos
i sin
6u& & &

i
"
& " $6u& & " &
" @&
" "
"
"
% !% " " % % " % u% 1& %
p t
cos cos sin i sin sin r cos sin cos p0 r sin cos i sin cos cos i sin p0 sin cos r
(7.5) (7.6) (7.7) (7.8) (7.9) (7.10)
(7.11) (7.12)
(7.13) (7.14)
(7.15) (7.16)
% 6& P %
(7.17) (7.18)
(7.19) (7.20)
56
monochromatic so
The total power is then
0 1 1
E XAMPLE 7.2
R OTATING MULTIPOLE Two point charges of equal charge q are located in the xy plane at either end of the diameter of a circle of radius a. The particles rotate with a constant angular speed in the plane of the circle. Determine
(a) The Fourier components of p1 , m, and Q

and
Solution
(b) The radiation diagram when a
w&
" %
dx 1
x2
6&
" %
b F
0 p2 4 0 2 32 2 c 0 p2 4 0 2 32 2 c 0 p2 4 0 6 c
F (
Sr r2 d
d sin 1
&
" %
0 p2 4 0 1 32 2 cr2
cos2 r
cos2
& 6u&
&
kk
E ND
6&
&
&
% u%
% ! &
% &
% %
1 1 0 1 1 E B 2 0 20 4 40 r2 kp k p k 32 2 0 r2 p k 2r 32 2 0 r2
(b) We use the formula S
1 20 E
B where the E and B elds are complex and

(7.21) (7.22) (7.23) (7.24)
(7.25) (7.26) (7.27) (7.28)

OF EXAMPLE
7.1
7.3. S OLVED
EXAMPLES
57
Fourier transform
kh &
! %
kh & %
" n
qa2
1 0 0 0 1 0 0 0 0
qa2 20
1 0 0 0 1 0 0 0 0
&
b F
Qxy
qa2
e b
d b
1 2
i 20 t
dt
& u&
20
qa2 i
20
&
! % b F
! & % %
Qyy
qa2
1 2e
1 1 2 1 qa2 2
1 2ei20 t
i20 t
ei t dt
& u&
" @&
&
"
" %
" & % %
b F
Qxx
1 2 qa 1 cos 20t ei t dt 2 qa2 1 2 20 1 2
&
&
! % " %
Qxx Qyy Qxy
qa2 1 e i20 t qa2 1 e i20 t Qyx qa2 ie i20 t
2 0
&
" %
&
! %
6&
"
Qxx Qyy Qxy Qiz
q x2 x2 2qa2 cos2 0t qa2 1 1 2 2qa2 sin2 0t qa2 1 cos 20t Qyx qa2 sin 20t Qzi 0
T T
Qi j
qn xin x jn
cos 20t
& %
b F
z q0 a2
1 2
e i t d
q0 a2 z
& 6u& T
!%
&T !% "T
T %
T
r1
1 r qvn 2 n n
T
qn rn
0 1 r qv1 2 1 q v1 q0 a2 z
&
&
"
"
!%
T ! T T T ! T
T
r1 r2 v1 v2
a cos 0t x sin 0t y r1 0 a sin t x cos 0t y v1
(7.29) (7.30) (7.31) (7.32) (7.33) (7.34) (7.35) (7.36) (7.37) (7.38) (7.39) (7.40) (7.41) (7.42) (7.43)
(7.44) (7.45)
(7.46) (7.47)
(7.48)
(7.49)
58
i0 eikr k Q 8 r i eikr k Q 80 r
E XAMPLE 7.3
ATOMIC RADIATION A transition in an atom is described as quantum matrix element of a radiation operator between the un-normalised eigenstates
At a certain moment, the atom is therefore described by where cf and ce can be viewed as given constants, chosen such that becomes normalised. According to semiclassical theory, one can interpret the magnitude squared of the wave function as a particle density function. Determine, according to this semiclassical interpretation, the power emitted by the atom via the dipole radiation which appears due to the transition between the two states. The power from an electric dipole is given by
Solution
The charge density is
Only the two last terms contribute because they are non-static. Via inspection we nd the Fourier components for E1 E0
&
!%
6& P %
qc f ce r cos exp
2r
&
!%
% %
" & &
u&
%
% !%
c f ce exp
i E1
E0 t
c f ce exp i E1
E0 t
y& &
" u&
!%
0"
&
!%
c f 2 r2 exp
2r cos2
ce 2 exp
2r
r cos exp
0"
"
"
w "
0 4 p 2 12 c
"
cf f
ce e
& !% & !%
f e
re r exp i 1 t cos E e r exp i 0 t

E
(7.53)
(7.54)
(7.55)
cf f
c f ce e f
c f ce e f
ce e
(7.56) (7.57) 2r (7.58)
(7.59)
&
1 E 2 0
&
&
% u%
&
k k
(7.50) (7.51)
(7.52)
E ND
OF EXAMPLE
7.2
7.3. S OLVED
EXAMPLES
59
0 2 0
and a similar integral for the complex conjugate term. The power from the electric dipole radiation due to the transition between the two states is given by
C LASSICAL P OSITRONIUM Calculate the radiation from a positron-electron head-on collision and subsequent annihic up lation, classically by assuming that the particles travel at a constant velocity v0 until the time the annihilate.
Solution
Background Consider a system of localised charges in motion. If we assume that we R much greater than extension x of the charge system are observing at a distance x and further that v c, it can be argued that the source time t is approximately t R c (where R x ) instead of t x x c. This is because, due to the rst assumption x (7.70) x x R x R Since the timescale for the system is of the order T x v and since v c we have that
x c
(7.71)
T
A &
T
! !
f
&
0 4 p 2 12 c
16 q2 0 E1 3c 4
E0
& 6u&
%
& P P % " P P %
12 qc f ce 0 0 1 8 qc f ce 0 0 1 8 qc f ce z
1 21 3
c f ce
6&
&
F
" P P %
12 qc f ce 0 0 1
1 2
sin 3
sin d
a&
&
"
%
12qc f ce 0 0 2
1 2 sin
sin cos 2 d
a&
FP P %
12qc f ce 0 0
cos2 sin d d
E ND
OF EXAMPLE
&
P P %
1 qc f ce 4! 2
cos sin sin cos sin sin cos d d
F )
&
!%
qc f ce
r4 exp
2 0
2r dr
cos sin r d d
&
!%
'
F (
qc f ce r cos exp
2r r r r2 dr d
Consequently, the corresponding Fourier component of the dipole moment p is
r d3x
(7.60) (7.61) (7.62) (7.63) (7.64) (7.65) (7.66) (7.67) (7.68)
(7.69)
7.3
E XAMPLE 7.4
60
and this allows us to write
The vector potential is this case 0 A j t R c d3x (7.73) 4 R since the denominator in the integrand is now not dependent on the source coordinates. Substituting j v, we rewrite the vector potential as 0 A qv (7.74) 4 R Observe that the summation can be written as d qx p (7.75) qv dt where p is the electric dipole. Thus, 0 A p (7.76) 4 R Deriving the EM eld in the usual manner we get
0 R p 4 rc2 R It can be shown that the angular and spectral distribution of energy is
The Calculation The dipole electric moment of the positronium is
We take the second time derivative of this p 2qe v0 t z t
2p 2qe v0 t z t2 For the spectral and angular distributions of the radiation we have
d p R R 2 (7.83) d d 2 c3 where subscript denotes the Fourier transform of the dipole. Now the Fourier transform of the dipole is simply p 2qv0 z 2 1 2 , so that
&
%
&
!
& %
& !%
!
2qe v0t z 0
t t
&
% & u%
d d d
R R 2 c3
&
0 0
&
&
% u%
1 p 40 rc2
R R
&
! %
&
T
R c
(7.72)
R R
(7.77) (7.78)
(7.79)
(7.80)
(7.81) (7.82)
7.3. S OLVED
EXAMPLES
61
d q2 v2 z R 2 2 c3 (7.84) e 0 d d In the nal result we notice that there is no dependence on so the spectral density is white noise.
&
E ND
OF EXAMPLE
7.4
62
L ESSON 8
Radiation from Moving Point Charges

8.1 Coverage
y z
In our series on deriving elds from given sources we have come to the most fundamental case: the moving point charge. The elds are derived from the LinardWiechert potentials. In what follows we will assume that the motion x t is known in advance. We nd that accelerating charges radiate. We also look other mechanisms for a point charge to radiate such as Cerenkov emission.
8.2
Formulae used
According to the Formulae (F.2326), the elds from a charge in arbitrary motion are given by q Rv 1 40 s3 v2 c2 Rv v 2 c
v c
y
y
Bt x
x x
Et x cx x
y
y z
Et x
z y z y
(8.1)
(8.2)
(8.3)
63
64
FROM
M OVING P OINT C HARGES
Alternative formulae
8.3
E XAMPLE 8.1
Solved examples
P OYNTING VECTOR FROM A CHARGE IN UNIFORM MOTION Determine the Poynting vector for the elds from a charge q which moves with constant velocity v. Show that no power is emitted from the charge during the motion.
Solution
In general the elds due to a single point charge may be written as
"
"
E B
Ev Bv
Erad Brad
t t
s r
cos
y
r r
r r cos
Rv
x x0
r r
v c
t t
y
Rv
v c
0 z
(8.4)
(8.5)
(8.6)
v c
(8.7)
rv
r0
(8.8)
r 1 cos
(8.9)
(8.10)
(8.11) (8.12)
8.3. S OLVED
EXAMPLES
65
Ev and Bv are known as the velocity elds and Erad and Brad are known as acceleration elds. These can be derived from the Linard-Wiechert potentials and result in
and
The big outer square brackets indicate that one should evaluate their content at the retarded time t t x x t c, where x t is the given motion of the charged particle. It is this that makes the equations for the elds difcult to evaluate in general. It may not be difcult to get an expression for x t , but then to solve the equation for the retarded time t t x x t c for t t t x . In this case it is not necessary to perform this transformation of variables since we are not interested in the time evolution, so we drop the brackets. It is easy to verify that for uniform motion of the charge q, or in other words 0, that
Furthermore, it can be shown that
where is the angle between r0 and . Inserting this relation for E into the relation for S we obtain
&
! %
&
! %
q 1 2 3 40 r0 1
1 2 sin2
r 3 2 0
where we have introduced the virtual position vector r0 we may write
&
% !
6&
% !
2 r0
r0
&
6&
1 E 0 c
Erad
0 and that 1 B E c so the Poynting vector in this case is
Brad
0 E 0
r . With these last relations (8.21)
& P %T T & hT % T
& hT % T
x&hT % T yw
x& T % T
! !
! !
&
T
T
% !
where
v c and (8.17)
&
! %w
Brad
Bv
q r 1 40 cs3 q r r 40 c2 s3
&
! %w
Erad
&

! % &
! %
Ev
q r r 1 2 40 s3 q r r r 40 cs3
(8.13) (8.14)
(8.15) (8.16)
(8.18)
(8.19)
(8.20)
66
FROM
So
and
So nally S
And now when we integrate the Poynting vector over a spherical surface A with radius R which encloses the moving charge which results in
since the Poynting vector is not radial, so a charge in uniform motion in vacuo, does not radiate energy.
E XAMPLE 8.2
S YNCHROTRON RADIATION PERPENDICULAR TO THE ACCELERATION Determine the angular distribution of synchrotron radiation in the plane perpendicular to the acceleration v for a charged particle moving with velocity v.
Solution
We consider only the formulas for the radiation elds, for which the denominator is the cube of the retarded relative distance
where is the angle between the velocity and r. So that

&
v r r r v r2 r2 cos vr2 1 cos vrs
&
&
&
!
Q&
rv
rv r v
v r rv
! % ! ! % ! ! % ! 4 r ! & t sq % p
Now, we have that v r
0, so
&
! %
r r
r r cos
r 1 cos
(8.27)
(8.28a) (8.28b) (8.28c) (8.28d)
&
S dA
S r d
&
&
! % ! %
q2 v1 4 16 2 0 r0 1
! w&
!
r0
r0
r0
r0 sin
2 r0 sin
2 2 sin 2 sin2 3
&
E ND
&
! %
&
! %

r 3 0
r0
0 q2 1 2 2 6 2 0 16 2 0 r0 1
1 2 sin2
(8.22)
(8.23)
(8.24)
(8.25)
(8.26)
OF EXAMPLE
8.1
8.3. S OLVED
EXAMPLES
67
0 q 0 q vr vrs 4 s3 4 s2 The Poynting vector is given by
Remember that the Poynting vector represents energy ux per unit time at the eld point at the time t. The charges energy loss must be related to the time t , i.e., the time when the energy was emitted!
but
T HE L ARMOR
FORMULA
Derive the Larmor formula by calculating the radiated power of an accelerating point charge due to electric dipole emission. Apply the Larmor formula to linear harmonic motion and circular motion.
Solution
The Larmor formula is a very useful equation for deriving the power of emission from non-relativistic accelerating charged particles. It can be derived from the radiation elds of a non-relativistic (v c) charged particle, but it can also be seen as an electric dipole relative a co-moving coordinate system. We shall investigate the latter. First derive the power emitted by an electric dipole. Using a time domain (non-Fourier) version of the dipole elds:
0 p r 4 rc we nd the Poynting vector to be
&
&
1 p 40 rc2
(8.33) (8.34)
&
! %
&
&
! %
! %
U t
U t
Sr r2
0 q2 t U 1 cos t t 16 2 c r2 1 0 q2 v2 1 16 2 c 1 cos 3
v2 cos
r2 (8.32)
&
! %
1 E2r 0 c
!
!
Erad
(8.29)
2 v2 r 0 q2 2 c r2 1 16 0 cos
(8.30)
(8.31)
E ND
OF EXAMPLE
8.2
E XAMPLE 8.3
68
FROM
0 2
p 2 sin3 d d
p2 4 80 c3 3 p2 60 c3
Linear harmonic motion

q
2 4 0 x 2 0
Circular motion
P&
6& % P &
6& %
xt
6& %
Pt
cos 0t 60 c3
2
4 q2 0 x2 0 120 c3
R0 cos 0t
yt
R0 sin 0t
A &
! 6& %
& % 6& % P &
6& %
xt
x0 cos 0t
at
xt
2 0 x0 cos 0t
& %
8 6& %
where we have identied the acceleration a t
6& %
Pt
q2 x 2 60 c3
& %
6& %
pt
qx t
p2 80 c3
sin3 d
q2 a 2 60 c3 xt .
1 16 2 0 c3 1 16 2 0 c3
F (Q& %
Pt
S r r2 sin d d
& %
So integrating for the power P t
r 2 sin d d
&
% &
6&
r (8.35) (8.36) (8.37) (8.38) (8.39) (8.40) (8.41) (8.42)
&

1 p r p r r 16 2 0 r2 c3 1 p r 2r r p r 2 r 2 c3 16 0 1 p r 2r 2 r 2 c3 16 0
1 E 0
8.3. S OLVED
EXAMPLES
69
Validity of the Larmor formula The Larmor formula although not covariant in form
can indeed be extended such as to be valid for all inertial frames. One point that should denitely be raised is power radiated for more than one accelerating charge. It is not so simple that one may assume that power is proportional to the number of sources N. What must be understood is whether or not the sources are radiating coherently or incoherently. For example, consider the above case of circular motion. If we had a large collection of particles such as the case with in a particle storage ring or in circular wire, the radiation is not automatically proportional to N. If they are bunched the power is proportional to N 2 , this is coherent radiation. If the charges are distributed homogeneously the radiated power is 0. And if the charges are distributed evenly but uctuate thermally then power is proportional to N, this is incoherent radiation.
VAVILOV- C ERENKOV
EMISSION
Show that the potentials at time t at a point inside the Vavilov-Cerenkov cone receive contributions from exactly two positions of the charged particle.
Solution
x2
v2 t
or, from the expression for r2 above:
This is a quadratic equation in t t . For a xed t we would in other words have two values of t . It remains to show that the result is physically reasonable.
"
" x& T ! %
& T
! %
" &T ! % )
v2
c2 n2
t t
2vx t t
&T ! %
" & T ! %
" &T ! %
x2 v2 t t
&T ! %
'
r2
c2 t n2
&T
! %
"
but r
c n
t is the retarded distance so that t

2
2vx t t
" @& T ! % " &T " & ! T % ! U
! % " &T
" ! %
r2
y2
vt
y2 (8.44)
2vx t
y2
y2
c2 t t n2
y2 x2
&
! T %
& 6hT % T
The motion of the charge is given by x t
vt
t so that
(8.45)
0 0
(8.46) (8.47)
w& %
Pt
4 q2 0 R2 0 60 c3
(8.43)
E ND
OF EXAMPLE
8.3
E XAMPLE 8.4
70
FROM
where c is the critical angle of the Cerenkov radiation; it is half of the opening angle of the shock wave of the radiation. So that, in other words, x y is inside the cone! It remains to be shown that x 0
tC
r sq
Bp
t )
r sq
&T ! %
t t
v2
c2 n2
vx
cos2 c cos2
vx
cos2 c cos2
E ND
OF EXAMPLE
vx
&
! % ! &
! %
cos2 c 1 cos2
cos2
vx
sin2 c
1 cos2
tan2
vx
&
c2 x2 y2 n2 v2 x2
! &
"
"
)
'
&T
! %
v2
c2 n2
c2 2 x n2
y2
v2 y2
y2 x2
& P %
R &
"
"
9 & T
'
! %
c2 2 x y2 v2 y2 n2 c2 y2 n2 v2 x2 y2 2 sin c sin2 c
v2
c2 n2
! &
"
vx
c2 n2
x2
y2
v2 y2 (8.48)
&
!
1& T
! %
vx vx 1
&
(8.49) (8.50) (8.51) (8.52) 1 (8.53) (8.54)
8.4
%&
"
2vx
4v2 x2 4 x2 y2 v2 2 v2 c2 n2
c2 n2
L ESSON 9
Radiation from Accelerated Particles

9.1 Coverage
The generation of EM elds via Linard-Wiechert potentials are considered simultaneously with the Lorentz force to give a self-consistent treatment of radiation problems. In the previous lesson we solve radiation problems from given expressions for the motion. Now we consider how charges actually move in the EM elds and thus present the Lorentz force. We also discuss the effect of radiation on the motion of the radiating body itself known as radiative reaction
9.2
y
Formulae used
qE v B
The covariant Lorentz force
d m v dt
(9.1)
71
72
FROM
ACCELERATED PARTICLES
9.3
E XAMPLE 9.1
Solved examples
M OTION OF CHARGED PARTICLES IN HOMOGENEOUS STATIC EM FIELDS Solve the equations of motion for a charged particle in a static homogenous electric eld E and magnetic eld B. Hint: Separate the motion v into motion parallel and perpendicular to the magnetic eld respectively.
Solution
Background on equations of motion for charged particles As we know, the motion of a charged particle in electric and magnetic elds is given by the Lorentz force
This can be seen as a denition of the E and B elds and also the fundamental equation for measuring the elds. The E eld gives the force parallel to direction of motion and B gives the force perpendicular to the direction of motion. In this certain sense the Lorentz force is trivial: it is simply a denition of the EM elds. The equation for the Lorentz force is relativistically correct as it stands, as long as one interpretes d m v (9.3) F dt where m is the mass of the particle, 1 1 v2 c2 , v is the three-velocity, and t is the time. The fact that the Lorentz force in this form d m v qE v B (9.4) dt is Lorentz invariant is not immediately clear but can easily be shown. On the other hand this equation is difcult to solve for v t because contains v. In many cases one has conditions which are non-relativistic and under such conditions it is possible to simplify (9.3) and thus also (9.4). One simply uses the fact that as v c 0 then 1. In this case F mdv dt, (i.e. the Newtonian force denition), so the Lorentz force becomes q dv E v B (9.5) dt m This equation is the equation of motion for non-relativistic charge particles in EM elds.
The motion As is well known, charged particles are uniformly accelerated in a static
and homogenous electric eld, and in a static and homogenous magnetic eld the charged particles perform circular motion. So what happens in a combined electric and magnetic elds? The equations of motion are given in (9.5) The rst step in solving this equation is to separate the motion into motion parallel with the B eld which we will denote with v and
& %
&
&
&
" %
" %
&
" %
&
qE
(9.2)
9.3. S OLVED
EXAMPLES
73
q E dt m and for the perpendicular direction
q E v B (9.8) m dv q q v E B (9.9) dt m m This is a rst order linear ordinary differential equation in v . It is also inhomogeneous which means that we have both a solution to the homogenous equation and solution to the inhomogeneous equation. Let us call the solution to the inhomogeneous equation D and the solution to the homogenous equation R . For the inhomogeneous solution, we notice that neither the inhomogeneity nor the coefcient of the zeroth order term depends on t so the solution D itself cannot be time dependent. With this assumption, and that is left is D B E . This is easily solved, by taking the cross product of this equation with B we nd
The solution to the homogenous equation
on the other hand, can be shown to have the form r c e i t where c is a constant vector perpendicular to B and satises c ic 0 (where the denotes scalar product dened as the inner product of the vectors), so for example c 1 i 0 if B is in the 3 direction.
Interpretation of the Motion Having derived the solutions for the motion of the charged particle we are now in a position to describe it in words. First of all we see that the motion consists of three separate parts. First we have the motion along the B eld which simply is not effected by the B eld, so that part of the E eld which is along the B eld accelerates the charge as if there were no B eld. The motion perpendicular to the B eld can further be separated into two parts. One part, which here we have denoted r , repreDraft version released 9th December 1999 at 19:47
& P P %
d r dt
q r m
B2
D B B E D B B B2 D
B E
&
"
" %
&
&t
4 r sq % p !
% !
dv dt
dv
&
" %
" x
dv dt dt and since the and
"
dv
q q E v B E m m are mutually orthogonal the equation separates into
" q
motion perpendicular to B, which we will denote v ; so v
v and thus (9.6)
(9.7)
(9.10) (9.11)
(9.12)
(9.13)
74
FROM
y x
Figure 9.1. Motion of a charge in an electric and a magnetic eld. Here the electric eld is along the y axis and the magnetic eld is along the z axis. There is a background velocity in the z axis for reasons of clarity. sents so called gyro-harmonic rotation, i.e. the particle moves in a circular orbit, with its axis of rotation parallel with the B eld with period of rotation 2 c where c qB m is known as the cyclotron frequency. This gyro-harmonic motion is the complete solution to our problem if the E eld was not considered. The second part of the motion in the perpendicular plane however, which we here denoted D involves both the B eld and the E eld. It is known as the drift velocity since this is the velocity of the centre of the gyro-harmonic motion. The motion is illustrated in Figure 9.1.
E XAMPLE 9.2
R ADIATIVE REACTION FORCE FROM CONSERVATION OF ENERGY Correct the equations of motion for a charged particle in EM elds to include the effects of the energy loss due to radiation emitted when the particle accelerates. Assume nonrelativistic conditions, so radiated energy is given by Larmor formula.
Solution
Background Among other things, Maxwells introduction of the electric and magnetic
elds was a mathematical technique to divide the work of solving the motion of charge particles in EM elds. Instead of action-at-a-distance, the elds naturally divide the prob-
E ND
OF EXAMPLE
9.1
9.3. S OLVED
EXAMPLES
75
lem into two parts: the generation of elds from moving charged particles and the motion of charged particles in EM elds. In a certain sense this can be seen as a division into, on the one side a cause and on the other side, an effect of EM interaction, but the formulas governing these processes are not very symmetric with respect to each other in classical electrodynamics: the generation of elds from moving point charges is determined via the Linard-Wiechert potentials and the motion of point charged particles is determined by the Lorentz force. Despite the lack of symmetry this division of seems successful at describing EM interaction except for one thing: it fails to describe the self-interaction of charged point particles. We know that motion of charged particles is governed by the Lorentz force but at the same time we know that acceleration of charged particles causes energy emission. These two facts have been until now treated separately, but taken together we realize something is missing. This can be seen for example in the case of a single charge under the inuence of a mechanical force in a region of space with no EM elds except for eld from charge itself. From the Linard-Wiechert potentials we know that the mechanical force will cause the charge to radiate and if energy is to be conserved the emission must take its energy from the kinetic energy but since the Lorentz force is zero and there are no other electromagnetic as of yet that we know of, we have no way of accounting for this radiative friction. This, as of yet not mentioned, force is known as radiative reaction or radiative damping force. One question that comes to mind after the above discussion is how can so many problems be described by classical electrodynamics without considering radiative reaction? Obviously it should have a negligible effect in most cases but what are the limiting conditions? Certainly, these should be determined by considering the conditions under which the energy emitted is of the same order as the kinetic energy of the charge. If we consider nonrelativistic motion, the energy emitted by an charge accelerating at the order of a, under a period of duration of order T , is given by the Larmor formula and is of the order
On the other hand, the acceleration bestows the charge with kinetic energy on th order of
or
q2 30 c3
&
q2 a2 T 60 c3
m aT 2
So if we demand that Erad equivalent to
Ekin
&
Erad
q2 a2 T 60 c3
(9.14)
m aT 2
(9.15) Ekin if we wish to neglect the radiation reaction, then this is
(9.16)
(9.17)
76
FROM
and if we dene the characteristic time as q2 30 c3 we can say that the effects of the . radiative reaction are negligible in measurements made over timescales on order of T
Accounting for radiative reaction Having demonstrated the need for a force which
accounts for radiative effects of accelerating charges we set out to determine its form. From the conservation of energy it is clear that the force we are looking for, which we will denote Frad , must satisfy
where have integrated the Larmor formula over time on the right hand side. Partial integration yields
if assume periodic motion then we nd that
and so
We now can correct the equation of motion to include the radiative reaction
This equation is know as Abraham-Lorentz equation of motion. Unfortunately, the Abraham-Lorentz equation of motion is not without its own inherent problems.
w&
! %
mv
Fext
Frad
q2 v 60 c3
m v
Frad
q2 v 60 c3
v dt
F
Frad v dt
q2 v v dt 60 c3
q2 v v 60 c3
F !
Frad v dt
q2 v v dt 60 c3
(9.18)
(9.19)
(9.20)
(9.21)
(9.22)
E ND
OF EXAMPLE
'
F F F
9.2
9.3. S OLVED
EXAMPLES
77
An electron moves in a circular orbit in a synchrotron under the action of the homogeneous magnetic eld B0 .
(a) Calculate the energy which is lost in electromagnetic radiation per revolution if the
speed v of the electron is assumed to be constant.
(b) At which particle energy does the radiated energy per revolution become equal to the particles total energy, if B0 1 5 T?
Hint:
Solution
(a) Since we wish to consider synchrotron motion we use the relativistically correct equation of motion
dp qv B dt where the right hand side is the Lorentz force with E momentum 4-vector so dp d d m0 v m0 v m0 v dt dt dt where
dp m0 v (9.27) dt This means that the power necessary to keep the particle in a circular orbit, disregarding radiation losses, is
which comes as no surprise. However, the circular motion does emit EM radiation. To nd the expression for the power loss due to radiation we need an expression for the accelera-
& )
"
m0 v v
'
dp dt
2 2 v v v c2
But v v
0, since we know that the electron moves in a circular orbit, so
6&
1 3d v v 2c2 dt
3 v v c2
v2 c2
)
'
!
B T
d dt
d dt
1 3 2
d v2 dt c2
"
&
d m0 v dt
w&
%T
R ADIATION
AND PARTICLE ENERGY IN A SYNCHROTRON
E XAMPLE 9.3
dU dt
0 q2 v2 4 6 c
(9.23) 0, and p is the space part of the (9.24)
(9.25) (9.26)
d dt
0 and thus
(9.28)
78
FROM
tion v . To this end we use what is left of (9.23), that is
This equation is easily solved to give
The factor m 0 c is known as the synchrotron angular frequency and is the relativistic 0 value of the angular frequency for gyro-harmonic motion or cyclotron angular frequency. Now we can insert this expression into the relativistic generalisation of the Larmor formula
2 0 q2 v2 4 0 q2 c 4 2 dU v dt 6 c 6 c To nd the energy per revolution, we need the period of revolution which is T so 2 dU 2 0 q2 c 4 2 v dt c 6 c 0 q 3 3 B0 v2 3cm0 2 0 q2 c 4 2 v 6 c
(b) The task here is to equate the total energy and the radiated energy and solve for
velocity and then see what total energy that velocity is associated with.
So, after a little algebra,
Thus the particle energy for which the radiated energy is equal to the total particle energy is
"
"
v2 c2
c2 v2
c2 1 v2 2
0 q3 B0 3cm2 0
1
(9.35) (9.36) (9.37)
0 q3 B0 3cm2 0
1
0 q3 B0 3cm2 0
2cm2 0 0 q 3 B0
(9.38)
&
%
The total energy is E m0 c2 and the radiated energy is Urev which when equated gives
Urev
(9.33) (9.34)
0 q 3 3 v2 B0 3cm0 ,
qB
m0 v
q vB0
q B0 v m0
The acceleration is found by taking the norm of this last equation and since v have is the scalar equation
m0 v
qv
(9.29) B all we (9.30)
(9.31)
(9.32) 2 c ,
9.3. S OLVED
EXAMPLES
79
This is obviously the upper limit for which radiation effect can be neglected in the treatment of synchrotron motion of the particle.
R ADIATION LOSS OF AN ACCELERATED CHARGED PARTICLE A charged particle, initially at rest, is accelerated by a homogeneous electric eld E0 . Show that the radiation loss is negligible, even at relativistic speeds, compared to the particles own energy gain. Assume that under all circumstances E0 108 V/m.
Solution
The relativistic equation of motion is
Having found v we try to derive the radiation eld generated by this motion; we have that
&
0 q 1 r v r 4 r3 1 cos 3 0 q 1 r2 v sin 3 1 4 r cos 3 0 qv sin 4 r 1 cos 3
kh
4 n t sr q p
&
&
&
&
! %
! %
0 q 4 r3 1
1 cos
r v c
! w
&
&
! %
! %
! %
Erad
q E0 m0 3
0 q r rv v 4 s3 0 q 1 4 r3 1 cos
m0 3 v
q E0
)
m0 v
m0 v 2
q E0
"
6&
'
" %
m0 v 1
2 2
m0 v 1
'
Since v
E0 then
2 1 2
& )
"
m0 v
'
w&
%T
dp dt
d m0 v dt
2 vv v c2
qE0
(9.40)
(9.41) (9.42) (9.43)
v c
(9.44)
"
T
m0 c2
2cm2 0 0 q 3 B0
0 16 TeV
(9.39)
E ND
OF EXAMPLE
9.3
E XAMPLE 9.4
80
FROM
where have used the particular geometry of the vectors involved. Then we may determine the Poynting vector, which is
The radiated energy per unit area per unit time corresponding to this is Sr r
2
but the energy radiated per unit area per unit time at the source point is
The total radiated energy per unit time is U t
From the equation of motion we nd that
Hence
m0 v
q E0 2
d d m0 v m0 v m0 v dt dt and, from the expression for v,
"
w&
dE dt
m0 c 2
d dt
q E0
We compare this expression for the radiated energy with the total energy E
0 1 x2 8 2 c 1 x 5 1 x 5 1 2 2 qv 4 q2 v2 4 6 1 2 0 2 2 0 2 2 3 8 c 3 1 8 c 3
&
" %
&
&
" %
! %
b F
dx
2 0 q4 E0 6 m2 c 0
&
" %
'
dx sin q2 v2 1
b F
!
x d
cos
&
! %
U d t
0 q2 v2 2 16 2 c
&
0 q2 v2 sin2 2c 1 16 cos
sin3 d 1 cos 5 0 q2 v2 1 1 x2 2 0 2 dx 8 c x 5 1 1
&
! %
&
! %
! %
r2 1
2 cos 0 q2 v2 sin2 0 c 16 2 r2 1 cos
#
F
U t
U t t t
U s t r
0 c
U t
Erad

2
%T
1 2 Erad r 0 c
Sr r
(9.45)
r2
(9.46)
rs Erad 0 c
(9.47)
(9.48)
m0 c2 , so (9.49)
(9.50)
(9.51)
9.3. S OLVED
EXAMPLES
81
v2 c2
and
31 2
11
10
E ND
OF EXAMPLE
Which of course is a very small ratio for a relativistic electron (v even for E 108 V/m.
2 998
&
&
E0 4 10 7 1 6 10 19 v 6 2 998 108 9 11 10
A % A % & b
%
dU dt
dE dt
2 0 q4 E0 1 6 cm2 q E0 v 0
0 e 3 E0 6 cm2 v 0
3 12 E0
Finally, for E0
108 V/m, and an electron for which q
m0 c 2
dE dt
d dt
q E0 v m0 c 2 q E0 v m0 c 2
q E0 v e, we nd that
t
(9.53) (9.54) v (9.55) 108 m/s)
9.4
r sq
m0 v
q E0 1
'
d dt
1 2
(9.52)
82
A PPENDIX F
Formulae
F.1 The Electromagnetic Field
F.1.1 Maxwells equations

P D B E H
0 B t j D t
(F.1) (F.2) (F.3) (F.4)
Constitutive relations
D
E B
E 0 E

(F.5) (F.6)
(F.7) (F.8)
83
84
A PPENDIX F. F ORMULAE
F.1.2 Fields and potentials

Vector and scalar potentials
Lorentz gauge condition in vacuum
F.1.3 Force and energy

Poyntings vector
Maxwells stress tensor
F.2
Electromagnetic Radiation
k
F.2.1 Relationship between the eld vectors in a plane wave
E c
F.2.2 The far elds from an extended source distribution

V
z y
Erad x
d3 x e
ik x
j k
i eik x x 40 c x
z y
Brad x
d3 x e
ik x
i0 eik x 4 x
Ti j
Ei D j Hi B j
1 c2 t

(F.9)
A t
(F.10)
(F.11)
(F.12)
1 ED 2 ij k k
Hk Bk
(F.13)
(F.14)
(F.15)
(F.16)
F.2. E LECTROMAGNETIC R ADIATION
85
F.2.3 The far elds from an electric dipole

$ z y
x rad B x
F.2.4 The far elds from a magnetic dipole

$ z y
rad B x
F.2.5 The far elds from an electric quadrupole

z y
rad B x
F.2.6 The elds from a point charge in arbitrary motion

y
Et x x x
t t
y
Rv
z
v c v c
y
y
Bt x
x x
Et x cx x
q Rv 1 40 s3
v2 c2
Rv v 2 c
y z
z y
Erad x
k Q
i eik x 80 x
z y
Erad x
i0 eik x k Q 8 x
k eik x m 40 c x
$ z y
Erad x
0 eik x m 4 x
z
k k
1 eik x p 40 x
0 eik x p 4 x
(F.17)
(F.18)
(F.19)
(F.20)
(F.21)
(F.22)
z y z y
(F.23)
(F.24)
(F.25)
(F.26)
(F.27)
86
F.2.7 The elds from a point charge in uniform motion

0 0 0 1
Bt x
Et x c2 R0
F.3
Special Relativity
1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1
F.3.1 Metric tensor

g
F.3.2 Covariant and contravariant four-vectors
v g v (F.33)
F.3.3 Lorentz transformation of a four-vector

x x (F.34)
u c
1 1
0 0
0 0
0 0 1 0
R0
x0
R0

v
y z
2
Et x
q 40 s3
v2 c2
z y z y
R0
(F.28)
(F.29)
(F.30)
(F.31)
(F.32)
(F.35)
(F.36)
(F.37)
F.4. V ECTOR R ELATIONS
87
F.3.4 Invariant line element
def
F.3.5 Four-velocity

u dx ds
F.3.6 Four-momentum
E cp
F.3.7 Four-current density

j
0 u
v c
F.3.8 Four-potential
z y
A
cA
F.3.9 Field tensor

F
F.4
Vector Relations
z y z y z y z y
Let x be the radius vector (coordinate vector), from the origin to the point x1 x2 x3 x y z and let x denote the magnitude (length) of x. Let further x x be arbitrary scalar elds and a x b x c x d x arbitrary vector elds. The differential vector operator is in Cartesian coordinates given by xi

i 1
xi
def
x, x 2 y, and x 3

where x i , i
1 2 3 is the ith unit vector and x 1
A x
A x
0 Ex Ey Ez
Ex 0 cBz cBy
Ey cBz 0 cBx
Ez cBy cBx 0
y z
y z y z y z
ds
dt
c d
(F.38)
v c
(F.39)
m0 c2 u
(F.40)
(F.41)
(F.42)
(F.43)
(F.44) z . In
88
component (tensor) notation
F.4.1 Spherical polar coordinates

Base vectors
sin cos x sin sin y cos z
Directed line element
Solid angle element
Directed area element
Volume element
d3x
dV
drdS
d2xn
sin d d
dS
dS r
r 2 d r
r2 dr d
dx x
dl
dr r

sin x
cos y
r d r2 sin d
cos cos x
cos sin y sin z
x1 x2 x3
can be written
x y z
(F.45)
(F.46)
(F.47)
(F.48)
(F.49)
(F.50)
(F.51)
(F.52)
89
F.4.2 Vector formulae

General relations
a b
z
a a
a b
c d
a b dc
a b cd
y z
a b
c d
a b
c d
ac bd
ad bc
y y y z z
b c
c a
a b
z
z
b c
ba c
ca b
y z
a b
b c

a b
b a
i jk a j bk x i
a b
b a
i j ai b j
ab cos
(F.53)
(F.54)
(F.55)
(F.56)
(F.57)
(F.58)
y y
(F.59)
(F.60)
(F.61)
(F.62)
(F.63)
90
a b
Special relations
In the following k is an arbitrary constant vector.

1 x

x x3
x x
y z z
2 a
x x3 1 x
ab
4 x
y y z z
y y y z z G
y z E z
y z
y t
(F.64)
(F.65)
(F.66)
(F.67)
(F.68)
(F.69)
(F.70)
(F.71)
(F.72)
(F.73)
(F.74)
91
Integral relations
Let V S be the volume bounded by the closed surface S V . Denote the 3-dimensional volume element by d3x dV and the surface element, directed along the outward pointing surface normal unit vector n, by dS d 2x n . a d3x
dS a

S
If S C is an open surface bounded by the contour C S , whose line element is dl, then
a dl

dS
S
dl
dS
S
a d3x
dS
S
d3x
dS
k a
ka
z
$
y z
k x
k2
1 x
4 k x

if x
x x3

k x x3
z z z i y y y z y z y

k x
1 x
k x x3
(F.75)
(F.76)
(F.77)
(F.78)
(F.79)
(F.80)
(F.81)
(F.82)
(F.83)
92

EMFT Exercises

Uploaded by

Copyright:

Available Formats

EMFT Exercises

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EMFT Exercises

Uploaded by

Copyright:

Available Formats

Tobia Carozzi Anders Eriksson Bengt Lundborg Bo Thid Mattias Waldenvik

E LECTROMAGNETIC F IELD T HEORY E XERCISES

Draft version released 9th December 1999 at 19:47

Downloaded from http://www.plasma.uu.se/CED/Exercises

E LECTROMAGNETIC F IELD T HEORY

E LECTROMAGNETIC F IELD T HEORY Exercises

Preface 1 Maxwells Equations

Example 1.2 Maxwells equations in component form

Solution . . . . . . . . . . . . . . . . . . . . . . Example 1.3 The charge continuity equation . . . . Solution . . . . . . . . . . . . . . . . . . . . . .

2 Electromagnetic Potentials and Waves

Example 2.3 Fourier transform of Maxwells equations

Solution . . . . . . . . . . . . . . . . . . . . . . . Example 2.4 Simple dispersion relation . . . . . . . . Solution . . . . . . . . . . . . . . . . . . . . . . .

Draft version released 9th December 1999 at 19:47

4 Lagrangian and Hamiltonian Electrodynamics

5 Electromagnetic Energy, Momentum and Stress

6 Radiation from Extended Sources

41 41 42 Example 6.1 Instantaneous current in an innitely long conductor 42

8 Radiation from Moving Point Charges

9 Radiation from Accelerated Particles

Example 9.1 Motion of charged particles in homogeneous static EM elds . . . . . . . . . . . . . . . . . . . . .

Draft version released 9th December 1999 at 19:47

Draft version released 9th December 1999 at 19:47

Motion of a charge in an electric and a magnetic eld . . . . . . .

Draft version released 9th December 1999 at 19:47

6.1 6.2 6.3

Draft version released 9th December 1999 at 19:47

Draft version released 9th December 1999 at 19:47

Draft version released 9th December 1999 at 19:47

M ACROSCOPIC M AXWELL EQUATIONS The most fundamental form of Maxwells equations is

Draft version released 9th December 1999 at 19:47

(1.1a) (1.1b) (1.1c) 1 E c2 t (1.1d)

L ESSON 1. M AXWELL S E QUATIONS

known as the macroscopic Maxwell equations.

Draft version released 9th December 1999 at 19:47

which can be rewritten as

Now by introducing the D and the H elds such that

the Maxwell equations become

Draft version released 9th December 1999 at 19:47

(1.16) (1.17) (1.18)

(1.25) (1.26) QED

L ESSON 1. M AXWELL S E QUATIONS

M AXWELL S EQUATIONS IN COMPONENT FORM Express Maxwells equations in component form.

Draft version released 9th December 1999 at 19:47

(1.27) (1.28) (1.29) 1 E c2 t (1.30)

(1.31) (1.32) (1.33) (1.34) (1.35)

(1.36) (1.37) (1.38)

L ESSON 1. M AXWELL S E QUATIONS

2. Take the divergence of the Ampre-Maxwell law: B

Comparison shows that

Component notation In component notation, a derivation of the continuity equation

2. Take the divergence of the Ampre-Maxwell law:

Comparison shows that

Draft version released 9th December 1999 at 19:47

Use that the relation i jk Ai A j

Interpreting the continuity equation The equation

Draft version released 9th December 1999 at 19:47

Draft version released 9th December 1999 at 19:47

Electromagnetic Potentials and Waves

Consider the magnetic eld given in cylindrical coordinates,

Draft version released 9th December 1999 at 19:47

Q& & P % ! Q& ! & P % Q& & P % Q&

& P % ! & P % & P %

& P % 6& P % & P % F 4 r sq p & P % f F b c d b

P & P & P & P & P P & P

" " & ! % ! % % ! u%